Novel Lanthanum Zirconate-based Thermal Barrier Coatings for Energy Applications [1st ed.] 9783030586942, 9783030586959

This book describes the latest developments of lanthanum zirconate based thermal barrier coatings. The physical, thermal

369 46 4MB

English Pages X, 85 [92] Year 2021

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

Novel Lanthanum Zirconate-based Thermal Barrier Coatings for Energy Applications [1st ed.]
 9783030586942, 9783030586959

Table of contents :
Front Matter ....Pages i-x
Overview of Lanthanum Zirconate-Based Thermal Barrier Coatings (Xingye Guo, Zhe Lu, Yeon-Gil Jung, Jing Zhang)....Pages 1-11
Properties of Lanthanum Zirconate–Based Thermal Barrier Coatings (Xingye Guo, Zhe Lu, Yeon-Gil Jung, Jing Zhang)....Pages 13-27
Modeling of Mechanical Properties of Lanthanum Zirconate Crystal (Xingye Guo, Zhe Lu, Yeon-Gil Jung, Jing Zhang)....Pages 29-46
Modeling of Thermal Properties of Lanthanum Zirconate Crystal (Xingye Guo, Zhe Lu, Yeon-Gil Jung, Jing Zhang)....Pages 47-57
Modeling of Mechanical Properties of Zirconia Top Coat and Bond Coat Interface (Xingye Guo, Zhe Lu, Yeon-Gil Jung, Jing Zhang)....Pages 59-78
Concluding Remarks and Future Research Directions (Xingye Guo, Zhe Lu, Yeon-Gil Jung, Jing Zhang)....Pages 79-82
Back Matter ....Pages 83-85

Citation preview

Xingye Guo Zhe Lu Yeon-Gil Jung Jing Zhang

Novel Lanthanum Zirconate-based Thermal Barrier Coatings for Energy Applications

Novel Lanthanum Zirconate-based Thermal Barrier Coatings for Energy Applications

Xingye Guo • Zhe Lu Yeon-Gil Jung • Jing Zhang

Novel Lanthanum Zirconate-based Thermal Barrier Coatings for Energy Applications

Xingye Guo Faculty of Materials and Manufacturing Beijing University of Technology Beijing, China Yeon-Gil Jung School of Materials Science and Engineering Changwon National University Changwon, Republic of Korea

Zhe Lu School of Materials and Metallurgical Engineering University of Science and Technology Liaoning Anshan, Liaoning, China Jing Zhang Department of Mechanical and Energy Engineering Indiana Univ Purdue Univ Indianapolis Indianapolis, IN, USA

ISBN 978-3-030-58694-2    ISBN 978-3-030-58695-9 (eBook) https://doi.org/10.1007/978-3-030-58695-9 © Springer Nature Switzerland AG 2021 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Preface

Thermal barrier coatings (TBCs) are refractory materials deposited on gas turbine components, which provide thermal protection for metallic components at operating conditions. Recently, lanthanum zirconate has emerged as a promising candidate material for TBC applications, due to its lower thermal conductivity and higher thermal stability compared to traditional TBC systems, such as yttria-stabilized zirconia. This book describes the latest developments of lanthanum zirconate–based thermal barrier coatings. Chapter 1 summarizes the manufacturing techniques of lanthanum zirconate powder and thermal barrier coatings. Chapter 2 presents the physical, thermal, and mechanical properties of lanthanum zirconate powder and coatings. Processing and characterizations of lanthanum zirconate powder and coatings under various conditions are also examined. Theoretical studies on the powder and coating’s mechanical properties are presented in Chap. 3. Chapter 4 discusses the modeling work of the thermal properties of the powder and coatings. Chapter 5 focuses on the theoretical studies of the mechanical properties at the interface between the ceramic top coat and metallic substrate. Future research directions of lanthanum zirconate as the next-generation thermal barrier applications are proposed in Chap. 6. This work is the fruit of international scientific collaboration. Xingye Guo acknowledges the support by the National Natural Science Foundation of China (No. 51901006) and Scientific Research Program of Beijing Municipal Education Committee (KM202010005006). Zhe Lu acknowledges the support by the National Natural Science Foundation of China (No. 51702145). Yeon-Gil Jung acknowledges the financial support by the “Human Resources Program in Energy Technology” of the Korea Institute of Energy Technology Evaluation and Planning (KETEP), which is granted from the Ministry of Trade, Industry & Energy, Republic of Korea (No. 20194030202450), and by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIP) (2018R1A5A6075959). Jing Zhang acknowledges the support by the U.S. Department of Energy (No. DE-FE0008868).

v

vi

Preface

Any comments and suggestions about the work are greatly appreciated. Please send the message to Jing Zhang, [email protected]. Beijing, China Anshan, China Changwon, Republic of Korea Indianapolis, IN, USA

Xingye Guo Zhe Lu Yeon-Gil Jung Jing Zhang

Contents

1 Overview of Lanthanum Zirconate-Based Thermal Barrier Coatings������������������������������������������������������������������������������������������������������   1 1.1 Thermal Barrier Coatings ������������������������������������������������������������������   1 1.2 Synthesis of Lanthanum Zirconate Powder����������������������������������������   3 1.3 Crystal Structure of Lanthanum Zirconate ����������������������������������������   4 1.4 Deposition Techniques of Lanthanum Zirconate-Based Coatings����������������������������������������������������������������������������������������������   5 References����������������������������������������������������������������������������������������������������  10 2 Properties of Lanthanum Zirconate–Based Thermal Barrier Coatings������������������������������������������������������������������������������������������������������  13 2.1 Physical Properties������������������������������������������������������������������������������  13 2.1.1 Coating Density and Porosity ������������������������������������������������  13 2.1.2 Sintering Behavior������������������������������������������������������������������  14 2.1.3 Morphology of Coating Cracks and Pores������������������������������  15 2.2 Mechanical Properties������������������������������������������������������������������������  15 2.2.1 Elastic Modulus����������������������������������������������������������������������  15 2.2.2 Hardness and Fracture Toughness������������������������������������������  16 2.2.3 Erosion Resistance������������������������������������������������������������������  17 2.3 Thermal Properties������������������������������������������������������������������������������  18 2.3.1 Melting Point Temperature and Specific Heat Capacity��������  18 2.3.2 Thermal Conductivity ������������������������������������������������������������  20 2.3.3 Coefficient of Thermal Expansion������������������������������������������  22 2.3.4 Furnace Cycling Test (FCT) and Jet Engine Thermal Shock (JETS) Test������������������������������������������������������������������  24 References����������������������������������������������������������������������������������������������������  25 3 Modeling of Mechanical Properties of Lanthanum Zirconate Crystal��������������������������������������������������������������������������������������������������������  29 3.1 Motivation and Introduction ��������������������������������������������������������������  29 3.2 Theoretical LZ Model ������������������������������������������������������������������������  30 3.2.1 DFT Model�����������������������������������������������������������������������������  30 vii

viii

Contents

3.2.2 MD Model������������������������������������������������������������������������������  33 3.3 LZ Crystal Constant Optimization������������������������������������������������������  34 3.4 Stress-Strain Analysis and Anisotropic Elastic Moduli����������������������  36 3.4.1 Large Deformation Stress-Strain Analysis������������������������������  36 3.4.2 Small Deformation Elastic Moduli Calculation����������������������  37 3.4.3 Bader Charge Analysis and Charge Density Distribution ����������������������������������������������������������������������������  40 3.4.4 MD Tensile and Shear Simulations����������������������������������������  41 3.5 Summary ��������������������������������������������������������������������������������������������  43 References����������������������������������������������������������������������������������������������������  45 4 Modeling of Thermal Properties of Lanthanum Zirconate Crystal��������������������������������������������������������������������������������������������������������  47 4.1 Introduction of Thermal Properties Calculations��������������������������������  47 4.2 Thermodynamic Energy and Specific Heat of LZ������������������������������  48 4.3 Thermal Conductivity of LZ ��������������������������������������������������������������  50 4.3.1 Thermal Conductivity of Single Crystal LZ Using MD Simulation ����������������������������������������������������������������������  50 4.3.2 Thermal Conductivity of Polycrystalline LZ Coating By Using FE Method��������������������������������������������������������������  53 4.4 Summary ��������������������������������������������������������������������������������������������  55 References����������������������������������������������������������������������������������������������������  56 5 Modeling of Mechanical Properties of Zirconia Top Coat and Bond Coat Interface ��������������������������������������������������������������������������  59 5.1 Introduction of Interface Modeling����������������������������������������������������  59 5.2 DFT Methods of ZrO2/Ni Interface Model ����������������������������������������  60 5.2.1 ZrO2 (111)/Ni (111) Interface Model��������������������������������������  60 5.2.2 Adiabatic Work of Adhesion��������������������������������������������������  62 5.2.3 Stress-Strain Behaviors in Tensile and Shear Deformations����������������������������������������������������������  62 5.2.4 Bader Charge Analysis������������������������������������������������������������  62 5.3 DFT Interface Modeling Results and Discussion ������������������������������  62 5.3.1 Adiabatic Work of Adhesion��������������������������������������������������  62 5.3.2 Stress-Strain Behaviors in Tensile and Shear Deformations����������������������������������������������������������  63 5.3.3 Charge Density and Bader Charge Analyses��������������������������  69 5.4 MD Tensile and Shear Simulations of ZrO2/Ni Interface Model��������������������������������������������������������������������������������������������������  72 5.5 Summary ��������������������������������������������������������������������������������������������  76 References����������������������������������������������������������������������������������������������������  77 6 Concluding Remarks and Future Research Directions��������������������������  79 6.1 Concluding Remarks��������������������������������������������������������������������������  79 6.2 Outlook of Future Work����������������������������������������������������������������������  81 Index��������������������������������������������������������������������������������������������������������������������  83

Author Biographies

Xingye Guo  is an assistant professor in the Faculty of Materials and Manufacturing at Beijing University of Technology, China. He obtained his PhD in mechanical engineering from Purdue University. Dr. Guo is the recipient of the International Thermal Spray Association (ITSA) scholarship for thermal barrier coating studies in 2017. His research interests include thermal/environmental barrier coatings and multi-scale modeling. Dr. Guo’s research is supported by the National Natural Science Foundation of China and Scientific Research Program of Beijing Municipal Education Committee. Zhe Lu  is an assistant professor of materials and metallurgical engineering at the University of Science and Technology Liaoning, China. He received his PhD from Changwon National University. Dr. Zhe is active in the research fields of fracture mechanics in nano- and micro-structured materials, material design and fabrication of advanced structural ceramics, ceramic-matrix composites, and thermal and mechanical properties in thermal barrier coatings. His research is supported by the National Natural Science Foundation of China. Yeon-Gil  Jung  is a professor of materials science and engineering at Changwon National University. He received his PhD in inorganic materials science and engineering from Hanyang University. Professor Jung was a visiting scientist at NIST during 1997–1999 and 2003– 2004. He has an h-index of 29, an i10-index of 91, and his works have been cited 3678 times (Google Scholar, 2020). Professor Jung has contributed to over 330 publications and acts as a reviewer on 10 international journals. ix

x

Author Biographies

Jing  Zhang  is an associate professor of mechanical and energy engineering at Indiana University – Purdue University. He received his PhD in materials science and engineering from Drexel University. Professor Zhang’s research interests include high-temperature materials and additive manufacturing. His research has been funded by the U.S. DOE, NSF, NASA, AFRL, and the industry.

Chapter 1

Overview of Lanthanum Zirconate-Based Thermal Barrier Coatings

1.1  Thermal Barrier Coatings The development of gas turbine engines relates to a great variety of engineering disciplines including turbine design, combustion, and cutting-edge materials. The application of thermal barrier coatings (TBCs) on turbine components has dramatically increased the operation temperature and lifetime of the alloy components in gas turbine engine [1]. The engine efficiency of gas turbines is increased due to the increase of operation temperature [2]. TBCs were initially introduced to the gas-­ turbine industry in the mid-1970s [3]. Nowadays, TBCs are widely used in aeronautics, astronautics, motor industry, and heat power stations. They are extensively applied in high-temperature components of gas turbines to protect the surface of metallic parts, such as combustor wall, rotating blades, stationary guide vanes, blade outer air seals, and afterburners in the tail section of jet engines and so on, because, in general, superalloy materials used as substrate of high-temperature components allow a temperature below 1100 °C [4–7]. TBCs are multilayered coating systems deposited on the turbine components, especially the turbine blade, which can thermally insulate them and protect them against hot and corrosive gas stream [1, 6, 8]. The typical structure of TBCs includes four layers: (1) superalloy substrate, (2) bond coat, (3) thermally grown oxide (TGO), and (4) ceramic top coat. The bond coat consists of an MCrAlN intermetallic alloy with a thickness of 100–300 μm, where M is an element selected from nickel, cobalt, iron, and their mixture, and N is an element selected from yttrium, zirconium, hafnium, ytterbium, and mixture thereof [9]. Typically, the ceramic top coat can be deposited directly on the substrate using various techniques, such as air plasma spraying (APS), high-velocity oxygen-­ fuel (HVOF) spraying, vacuum plasma spraying, low-pressure plasma spraying, and diffusion deposition method [10–13]. In high-temperature operation conditions, the bond coat is inevitably oxidized and a TGO layer is formed at the interface between the bond and the top coats with a thickness of 1–10  μm. The main © Springer Nature Switzerland AG 2021 X. Guo et al., Novel Lanthanum Zirconate-based Thermal Barrier Coatings for Energy Applications, https://doi.org/10.1007/978-3-030-58695-9_1

1

2

1  Overview of Lanthanum Zirconate-Based Thermal Barrier Coatings

composition of TGO is α-alumina (α-Al2O3), which can work as a good oxygen diffusion barrier to protect the bond coat and substrate from further oxidation. On the other hand, the growth of the TGO layer can lead to TBC failure [14]. The ceramic top coat is one or multiple low-thermal conductivity ceramic layers with a typical thickness of 100–600 μm, which is deposited by the APS or electron-beam physical vapor deposition (EB-PVD) methods [5, 8]. The criteria for TBCs material selection include high melting point, low thermal conductivity, high CTEs, good thermal and chemical stability, no phase change, low sintering activity, good erosion resistance, and good foreign objective damage (FOD) or calcium–magnesium–alumino-­silicate (CMAS) resistance [15]. Currently, the state-of-the-art TBCs are 7–8wt % yttria (Y2O3)-stabilized zirconia (ZrO2) (8YSZ). The 8YSZ has a metastable tetragonal phase (t′), and Y2O3 is used to stabilize the ZrO2 structure. The 8YSZ has a relatively high melting point (2680 °C) [15], relatively low thermal conductivity about 2.0–2.3 W/m/K around 1000 °C for a fully dense status (0.9–1.2 W/m/K for 10–15% porosity) [16, 17], a relatively high CTEs (11 × 10−6 /K at 1000 °C) [15], and good thermal and chemical stability [18]. However, the maximum surface temperature that can be employed as TBCs of 8YSZ is limited to 1200 °C for long-time operation. At temperatures above 1200 °C, two important degradation mechanisms occur in 8YSZ. The first mechanism is that the t′ phase of YSZ will decompose to two equilibrium phases: tetragonal (t) and cubic (c) phases. During the cooling process, the t phase will transform to the monoclinic (m) phase, causing ~4% volume expansion. The second mechanism is the sintering of coating, which will change the microstructure as well as the mechanical and thermal properties. The phase change and sintering, and the associated thermal and mechanical property changes will finally lead to high thermally induced stress and shorten the coating’s lifetime [15]. However, some approaches can be chosen to enhance TBCs’ performance: (1) find a new feedstock powder with a higher sintering resistance or no phase change and (2) control the porosity of TBCs [19]. In addition, increasing the thickness of top coat can reduce the substrate temperature, although the performance of TBC still remains the same. Lanthanum zirconate (La2Zr2O7, LZ) is a typical pyrochlore structure ceramic material. The general chemical formula of the pyrochlore structure is A2B2O7. The A element in A2B2O7 generally is a rare earth or an element with an inert single pair of electrons and the B element typically is a transition metal or a posttransition metal with a variable oxidation state [10, 20]. Compared to 8YSZ, LZ is more advantageous in the application of TBCs: (1) no phase transformation from room temperature to its melting temperature is noted, (2) has a considerably high sintering resistance, (3) has a very low thermal conductivity (1.5–1.8 W/m/K at 1000 °C for a fully dense material), and (4) has a lower oxygen ion diffusivity, which protects the bond coat and the substrate from oxidation [15]. The principal drawback of LZ is its small CTE, which does not match the high CTE of the bond coat.

1.2  Synthesis of Lanthanum Zirconate Powder

3

1.2  Synthesis of Lanthanum Zirconate Powder There are several ways to synthesize LZ powder, which can be used in thermal spray process, including solid state reaction method, co-precipitation method, and the sol-­ gel method [15, 21–23]. LZ powder can be synthesized by solid state reaction method from a mixture of lanthanum oxide (La2O3, 99.9%) and ZrO2 (99%) powders at high temperatures (T = 1773 K) under an argon atmosphere for 10 h [21, 22]. The pure La2O3 is typically prepared by dissolving La2(CO3)3·8H2O in nitric acid and subsequently producing a precipitate by adding NH3. The precipitate is dried in air and then heated in oxygen at 1173  K to remove the nitrogen-containing fragments. The La2O3 obtained is heated at 1473 K to remove any absorbed water [21]. The LZ powder thus obtained is spherical or ellipsoidal in shape with a porous microstructure on the surface. The theoretical chemical composition of LZ powder coating is La 48.6 wt%, Zr 31.9  wt%, O 19.5  wt%, which is equivalent to 1:2 molar ratios of La2O3 and ZrO2 [18]. In the co-precipitation method, an aqueous solution of lanthanum nitrate hexahydrate (La(NO3)3·6H2O) and zirconium oxychloride (ZrOCl2·8H2O) with a diluted NH3 solution is used to prepare LZ powder [15]. During this preparation process, La(NO3)3·6H2O and ZrOCl2·8H2O are dissolved in distilled water in equimolar amounts. The liquid mixtures are slowly added under stirring to an ammonium hydrate solution at pH 12.5. The precipitate obtained is filtered, washed with distilled water, and then dried at 120 °C overnight. The remaining solid is then calcined at 900 °C for 5 h. The sol-gel method is another way to synthesize nanostructured powders, which have high sintering ability for the A2Zr2O7 (A = La, Nd, Sm) system [23]. Tang et al. fabricated LZ fibers by a sol-gel combined electrospinning method [24]. LaCl3·7H2O (9  mmol) was dissolved in zirconium acetate solution in an equimolar ratio and 0.4 g silica sol (SiO2, 22%, pH = 3.2) was added with magnetic stirring for 6 h. The transparent sol was added to absolute ethanol (8.0 ml) with polyvinylpyrrolidone (PVP, 0.4 g). The sol was transferred into a syringe and 20 kV power was supplied between the syringe tip and a stainless steel collector. The LZ gel fibers were fabricated on the collector under the electrostatic force, after the dry and calcination processes at high temperatures. In addition to the aforementioned three commonly used methods, other methods also can fabricate LZ powders. Tong et al. prepared LZ nanopowder using stearic acid combustion method [25]. La(NO3)3·nH2O and Zr(NO3)4·9H2O were used as the precursors, and stearic acid (C17H35COOH) was used as the solvent and dispersant. The mixed solution was ignited in the air, and the obtained powder was calcined at a series of increasing temperatures ranging from 600 to 900 °C for 5 h in the air.

4

1  Overview of Lanthanum Zirconate-Based Thermal Barrier Coatings

1.3  Crystal Structure of Lanthanum Zirconate LZ is a typical ZrO2-Ln2O3 system (Ln is lanthanide elements, Ln  =  La  →  Gd), which has a pyrochlore structure with space group Fd 3 m [26]. Although the compounds Ln2Zr2O7 (Ln = La → Gd) are stable at room temperature, an order-disorder transition occurs at high temperatures (>1500 °C). For example, the compounds get transferred from pyrochlore to defect fluorite structure, with the only exception of LZ (actually Ln2Zr2O7, Ln = Tb → Lu adopt the defect fluorite structure) [20]. The transition temperature depends on the radius of Ln ions (La, no transition; Nd, 2300 °C, Sm 2000 °C, and Gd 1530 °C) [27]. X-ray diffraction is widely used to identify crystal structures. Figure 1.1 shows the X-ray diffraction patterns of several Ln2Zr2O7 materials at room temperature [23]. Two peaks indexed (331) and (511) shown in the LZ pattern originate from the pyrochlore structure, which are also observed in Nd2Zr2O7, Sm2Zr2O7, and Gd2Zr2O7 patterns. Other peaks are commonly observed for pyrochlore and defect fluorite structures [23]. The Ln2Zr2O7 pyrochlore crystal is a cubic structure in space group Fd 3 m (origin choice 2) with four crystallographically independent atom sites (rare earth ion 1 1 1 1 1 Ln, in 16d at ( , , ), Zr in 16c at (0,0,0), O1 in 48f at (x, , ), and O2 in 8b 8 8 2 2 2 3 3 3 at ( , , ). The structure type can be considered as an ordered defect fluorite 8 8 8 structure with the trivalent rare earth Ln3+ and quadrivalent Zr4+ cations forming an 7 ordered, face-centered cubic eutectic cation array. Oxygen ions are located in of 8 the tetrahedral interstices: O1 in an off-center position within the Ln2Zr2 tetrahedral and O2 in the Zr tetrahedral [28]. The x values of O1 can be varied from 0.3125 to 0.375. The x values 0.3125 and 0.375 correspond to the regular octahedral oxygen environment around the Zr4+ ion and regular cubic oxygen environment around Ln3+, respectively. Tabira et al. determined the x value of LZ using systematic row wide-angle convergent beam electron diffraction (CBED) techniques [28]. The results showed that the x value varied systematically with the rare earth ion radius, the larger radius corresponding to the larger x value. The experimental result of x value in LZ is 0.333, according to Tabira’s work [28]. The lattice parameter of a conventional cubic LZ cell can be calculated using XRD results based on Bragg’s law [29]. Shimamura et al. reported that the lattice parameters of Ln2Zr2O7 pyrochlores increased with the ionic radius of Ln, as shown in Fig. 1.2 [23]. The lattice parameter of LZ is 10.8 Å in Shimamura’s experiments and 10.802 Å in Tabira’s work [28]. As discussed from the XRD pattern in Fig. 1.1, the LZ has a cubic phase at room temperature. As shown in the ZrO2-La2O3 phase diagram of Fig. 1.3, the LZ has no phase transformation from room temperature to its melting point [30–32]. When the molar ratio of ZrO2 and La2O3 reaches 2:1, which corresponds to 33.3% La2O3, only a single LZ cubic phase is available from room temperature to its melting point.

1.4  Deposition Techniques of Lanthanum Zirconate-Based Coatings

5

Fig. 1.1  X-ray diffraction patterns of Ln2Zr2O7 [23]

1.4  Deposition Techniques of Lanthanum Zirconate-Based Coatings Air plasma spray (APS) is the most widely used thermal spray technique for LZ deposition. Plasma normally consists of neutral atoms, positive ions, and free electrons. Plasma is produced by transferring energy into the gas until the energy level

6

1  Overview of Lanthanum Zirconate-Based Thermal Barrier Coatings

Fig. 1.2  Lattice parameters of Ln2Zr2O7 as a function of ionic radius of Ln. Solid and open symbols stand for pyrochlore and defect fluorite structures, respectively [23]

is large enough to ionize the gas, allowing the electrons and ions to move independently of one another. As a result, plasma is often called the fourth state of matter. The plasma state is achieved when the current can be sustained under an electric field and the free electron can move through the ionized gas [10]. The plasma temperature in the core of the heating gas exceeds 20,000 °C, which depends on the gas properties. During the APS process, feedstock powders are carried by some noble gasses, such as argon and nitrogen, to the APS torch. The thermal plasma can be generated using electric arcs. Natural air can be used as the source of the plasma gas for the LZ spray. The essential APS parameters include current, carrier air flow rate, primary air flow rate, spray distance, powder feed rate, substrate tangential speed, etc. [33]. As shown in Fig.  1.4, the APS-sprayed LZ has many amorphous pores and cracks, which is well known as the “splat” grain [33]. The thermal conductivity of the APS deposited coating is lower than the EB-PVD deposited one, due to its “splat” grain morphology. The porosity of the LZ coating can be easily controlled by changing the spray parameters, which results in changing the thermal and mechanical properties. Another commonly used deposition technology is EB-PVD.  The term PVD denotes those vacuum deposition processes where the coating material is

1.4  Deposition Techniques of Lanthanum Zirconate-Based Coatings

7

Fig. 1.3  Phase diagram of ZrO2 and La2O3 in mole percentage [30–32]

evaporated by various mechanisms (resistant heating, high-energy ionized gas bombardment, or electron gun) under vacuum, and the vapor phase is transported to the substrate, forming a coating [34]. The main EB-PVD deposition parameters include vacuum pressure, substrate distance, power supply, average substrate temperature (1223  ±  25  K, 950  ±  25  °C), substrate rotate speed, etc. [35, 36]. As shown in Fig.  1.5, the EB-PVD-deposited LZ coating microstructure has a fine columnar microstructure that results in a higher strain tolerance [37]. Typically, the EB-PVD-­ deposited coatings have a higher thermal conductivity than the APS-sprayed coatings at the same porosity level. This is because the splat boundaries of the APS-deposited coatings act as scattering centers perpendicular to the direction of heat flux, which weakens the heat transfer process. However, the columnar grain boundaries in EB-PVD-deposited coating are parallel to the direction of heat flux. As a result, the EB-PVD-deposited coating has a good performance and considerably longer operating lifetimes [12].

8

1  Overview of Lanthanum Zirconate-Based Thermal Barrier Coatings

Fig. 1.4  Cross-sectional microstructure of APS-sprayed LZ coating [33]

Fig. 1.5  Cross-sectional microstructure of EB-PVD-deposited LZ coating [37]

However, the main disadvantages of EB-PVD technique are its low deposition rate, high investment costs, high thermal conductivity, and vapor pressure requirements. In addition to the two methods mentioned above, there are other techniques to deposit LZ coatings, such as suspension plasma spray (SPS) and spray pyrolysis [12, 38, 39].

1.4  Deposition Techniques of Lanthanum Zirconate-Based Coatings

9

Wang et al. deposited an LZ coating using SPS method [38] and a suspension of 30 wt.% nano-LZ particles in (99.5%) ethanol was produced. Meanwhile, an electrostatic dispersant of 1  wt.% polyethylene glycol (PEG1000) was added to the suspension. The suspension was injected into the plasma spray jet and the LZ coating was deposited. The main deposition parameters are standoff distance of the substrate and the concentration of the suspension feedstock. A liquid solution was used as a feedstock material instead of powder, which provided the possibility of tailoring the coating composition easily and facilitating the doped, multilayered, and graded coatings [39]. As shown in Fig. 1.6a, b, many pores and cracks were spotted in the microstructure of SPS-deposited coatings [38]. Weber et al. fabricated the LZ coating using spray pyrolysis method [39]. Zirconyl oxynitrate hydrate (ZrO(NO3)2·xH2O) and lanthanum nitrate hexahydrate (La(NO3)3·6H2O) were dissolved in deionized water and further mixed in a molar ratio of 1:1 of La to Zr. The precursor solution was sprayed at a flow rate of 1 ml/min at 240 °C. The deposited coatings were dried at 500–600  °C after the deposition process. As shown in Fig. 1.6c, d, a lot of cracks were observed in the coating layer [39].

Fig. 1.6  Cross-sectional microstructure of LZ coatings: (a) SPS deposited with standoff 40 mm, (b) SPS deposited with standoff 50 mm, and (c) surface view and (d) cross-section view of spray pyrolysis-deposited LZ coatings [38, 39]

10

1  Overview of Lanthanum Zirconate-Based Thermal Barrier Coatings

References 1. Clarke, D., & Levi, C. (2003). Materials design for the next generation thermal barrier coatings. Annual Review of Materials Research, 33(1), 383–417. 2. Perepezko, J. H. (2009). The hotter the engine, the better. Science, 326(5956), 1068–1069. 3. Miller, R.  A. (1997). Thermal barrier coatings for aircraft engines: History and directions. Journal of Thermal Spray Technology, 6(1), 35–42. 4. Miller, R. A. (1987). Current status of thermal barrier coatings — An overview. Surface and Coatings Technology, 30(1), 1–11. 5. Padture, N.  P., Gell, M., & Jordan, E.  H. (2002). Thermal barrier coatings for gas-turbine engine applications. Science, 296(5566), 280–284. 6. Clarke, D. R., Oechsner, M., & Padture, N. P. (2012). Thermal-barrier coatings for more efficient gas-turbine engines. MRS Bulletin, 37(10), 891–898. 7. Pan, W., et al. (2012). Low thermal conductivity oxides. MRS Bulletin, 37(10), 917–922. 8. Weber, S. B., et al. (2013). Lanthanum zirconate thermal barrier coatings deposited by spray pyrolysis. Surface and Coatings Technology, 227, 10–14. 9. Taylor, T.  A. (2011). Low thermal expansion bondcoats for thermal barrier coatings. US Patents No. 7910225 B2. 10. Davis, J. R. (2004). Handbook of thermal spray technology. Materials Park: ASM International. 11. Nelson, W. A., & Orenstein, R. M. (1997). TBC experience in land- based gas turbines. Journal of Thermal Spray Technology, 6(2), 176–180. 12. Fauchais, P. L., Heberlein, J. V., & Boulos, M. (2014). Thermal spray fundamentals: From powder to part. New York: Springer Science & Business Media. 13. Chen, M., et al. (2003). Characterization and modeling of a martensitic transformation in a platinum modified diffusion aluminide bond coat for thermal barrier coatings. Acta Materialia, 51(14), 4279–4294. 14. Sjöström, S., Brodin, H., & Jinnestrand, M. (2013). Thermomechanical fatigue life of a TBC– comparison of computed and measured behaviour of delamination cracks. In: ICF13. 15. Vassen, R., et al. (2000). Zirconates as new materials for thermal barrier coatings. Journal of the American Ceramic Society, 83(8), 2023–2028. 16. Schlichting, K. W., Padture, N. P., & Klemens, P. G. (2001). Thermal conductivity of dense and porous yttria-stabilized zirconia. Journal of Materials Science, 36(12), 3003–3010. 17. Hasselman, D. P. H., et al. (1987). Thermal diffusivity and conductivity of dense polycrystalline ZrO2 ceramics: A survey. American Ceramic Society Bulletin, 66(5), 799–806. 18. Cao, X. Q., et al. (2001). Thermal stability of lanthanum zirconate plasma-sprayed coating. Journal of the American Ceramic Society, 84(9), 2086–2090. 19. Lu, Z., et  al. (2014). Thermal stability and mechanical properties of thick thermal barrier coatings with vertical type cracks. Transactions of Nonferrous Metals Society of China, 24(Supplement 1), s29–s35. 20. Subramanian, M. A., Aravamudan, G., & Subba Rao, G. V. (1983). Oxide pyrochlores — A review. Progress in Solid State Chemistry, 15(2), 55–143. 21. Bolech, M., et al. (1997). The heat capacity and derived thermodynamic functions of La2Zr2O7 and Ce2Zr2O7 from 4 to 1000 K. Journal of Physics and Chemistry of Solids, 58(3), 433–439. 22. Sedmidubsky, D., Benes, O., & Konings, R. J. M. (2005). High temperature heat capacity of Nd2Zr2O7 and La2Zr2O7 pyrochlores. Journal of Chemical Thermodynamics, 37, 1098–1103. 23. Shimamura, K., et  al. (2007). Thermophysical properties of rare-earth-stabilized zirconia and zirconate pyrochlores as surrogates for actinide-doped zirconia. International Journal of Thermophysics, 28(3), 1074–1084. 24. Tang, H., et al. (2012). Fabrication and characterization of nanostructured La2Zr2O7 fibers. Material Letters, 70, 48–50. 25. Tong, Y., et al. (2008). Preparation and characterization of pyrochlore La2Zr2O7 nanocrystals by stearic acid method. Materials Letters, 62, 889–891.

References

11

26. Moriga, T., et  al. (1989). Crystal structure analyses of the pyrochlore and fluorite-type Zr2Gd2O7 and anti-phase domain structure. Solid State Ionics, 31(4), 319–328. 27. Michel, D., y Jorba, M. P., & Collongues, R. (1976). Study by Raman spectroscopy of order-­ disorder phenomena occurring in some binary oxides with fluorite-related structures. Journal of Raman Spectroscopy, 5(2), 163–180. 28. Tabira, Y., et al. (2000). Systematic structural change in selected rare earth oxide pyrochlores as determined by wide-angle CBED and a comparison with the results of atomistic computer simulation. Journal of Solid State Chemistry, 153(1), 16–25. 29. Bragg, W. H., & Bragg, W. L. (1913). The reflection of X-rays by crystals. Proceedings of the Royal Society of London. Series A, Containing Papers of a Mathematical and Physical Character, 88(605), 428–438. 30. Brown, F. H., & Duwez, P. O. L. (1955). The systems zirconia-lanthana and zirconia-neodymia. Journal of the American Ceramic Society, 38(3), 95–101. 31. Roth, R. S. (1956). Pyrochlore-type compounds containing double oxides of trivalent and tetravalent ions. Journal of Research of the National Bureau of Standards, 56(1), 17–25. 32. Maloney, M.  J. (2000). Thermal barrier coating systems and materials. US Patents No. 6,117,560. 33. Di Girolamo, G., et  al. (2015). Evolution of microstructural and mechanical properties of lanthanum zirconate thermal barrier coatings at high temperature. Surface and Coatings Technology, 268, 298–302. 34. Singh, J., et  al. (1999). An overview: Electron beam-physical vapor deposition technology-­ present and future applications. Applied Research Laboratory, Pennsylvania State University, USA. 35. Xu, Z., et  al. (2009). Double -ceramic-layer TBC of La2Zr2O7/YSZ deposited by electron beam-physical vapor deposition. Journal of Alloys and Compounds, 473, 509–515. 36. Saruhan, B., et  al. (2004). EB-PVD processing of pyrochlore-structured La2Zr2O7-based TBCs. Surface and Coatings Technology, 182(2–3), 175–183. 37. Bobzin, K., et al. (2013). Influence of temperature on phase stability and thermal conductivity of single- and double-ceramic-layer EB–PVD TBC top coats consisting of 7YSZ, Gd2Zr2O7 and La2Zr2O7. Surface and Coatings Technology, 237, 56–64. 38. Wang, C., et al. (2014). Nanocomposite lanthanum zirconate thermal barrier coating deposited by suspension plasma spray process. Journal of Thermal Spray Technology, 23(7), 1030–1036. 39. Weber, S. B., et al. (2011). Deposition mechanisms of thick lanthanum zirconate coatings by spray pyrolysis. Journal of the American Ceramic Society, 94(12), 4256–4262.

Chapter 2

Properties of Lanthanum Zirconate–Based Thermal Barrier Coatings

2.1  Physical Properties 2.1.1  Coating Density and Porosity The theoretical density of LZ material can be calculated using the molecular weight and the number of formula units per elementary cell [1]. Lehmann et al. calculated the theoretical density of LZ material as 6050 kg/m3 [1]. The porous LZ coating density can be measured following the ASTM standard B328-96, which is based on the Archimedes’ principle [2]. Freestanding samples, which were peeled off from the substrate without the bond coat, were used in the coating density measurements. Engine oil was used to seal the pores on the sample surface by dipping the sample with the oil, so the volume of the sample can be measured by Archimedes’ principle. Density (D, g/cm3) of the sample can be calculated by [2]: D

A D B C  E w

(2.1)

where A is the mass of oil-free specimen in the air, B is mass of oil-impregnated specimen, C is mass of oil-impregnated specimen and wire in water, E is the mass of wire in water, and Dw is density of water at specific immersion temperature. The porosity of the coating (φ) can be measured by the ratio of coating density to the theoretical density (Dt), which can be expressed by:



 D    1    100%  Dt 

© Springer Nature Switzerland AG 2021 X. Guo et al., Novel Lanthanum Zirconate-based Thermal Barrier Coatings for Energy Applications, https://doi.org/10.1007/978-3-030-58695-9_2

(2.2)

13

14

2  Properties of Lanthanum Zirconate–Based Thermal Barrier Coatings

The porosity of the LZ coating varies with the deposition parameters, such as deposition power, powder feed rate, and substrate standoff distance. For the APS-­ deposited coating, the higher the power used in the torch, and the larger the powder feed ratio and the shorter standoff distance used in the deposition process, the denser the coating will be. The commercial APS-deposited coating shows a porosity of 15–25%. However, it is important to tailor the porosity of the LZ coating in a range of 8–19% to acquire good thermal and mechanical properties [3].

2.1.2  Sintering Behavior Sintering in the porous ceramic coating is the process by which the coating will be densified by the reduction in surface energy associated with the excess surface area of the pores and cracks [4]. Sintering of the porous TBCs normally occurs at elevated temperature. When the sintering occurs, the densification process inevitably increases the elastic modulus and thereby decreases the strain compliance of the coating. Meanwhile, the thermal conductivity of the coating increases due to the decrease of the porosity [4]. Many experimental approaches can be used to describe the sintering behavior: (1) record the dimension change of the coating sample using a high-temperature dilatometer, (2) detect the porosity distribution change using mercury porosimetry, (3) measure the relative density of the coating during the sintering process, and (4) monitor the thermal conductivity increases during a long time heating period at various temperatures [5–7]. Vassen et al. investigated the sintering behavior of APS-­ deposited LZ coating at temperatures as high as 1650 °C using the high-temperature dilatometer and the mercury porosimetry [5, 6]. The higher porosity the coating had, the higher sintering rate it had. The coating with high porosity level (20%) densified strongly during the annealing process at 1250 °C [5]. Vassen et al. indicated that decrease porosity might lead to a better thermomechanical behavior [6]. Zhu investigated the sintering behavior by continuously monitoring the thermal conductivity [7]. Zhu found that the LZ coating showed significant thermal conductivity increase (from 0.55 W/m/K to 0.95 W/m/K in 20 h) at 1575 °C, suggesting the coating was undergoing substantial sintering. Nair et al. systematically studied the sintering of the LZ coating [8]. They found that the major mechanism of the sintering process in the temperature range of 800–1100 °C is surface diffusion. The sintering above 1100 °C is mainly because of grain boundary diffusion combined with the surface diffusion. The contribution from surface diffusion becomes negligible as the sintering temperature increases. In general, the sintering resistance of LZ coating is higher than that of YSZ coating, and also BaZrO3 and SiZrO3 coatings [5, 6]. The low-sintering activity of the LZ is beneficial for TBC applications.

2.2  Mechanical Properties

15

2.1.3  Morphology of Coating Cracks and Pores The crack and pore morphology of TBC is a crucial parameter affecting the thermal and mechanical properties of the coatings. The cracks can be categorized as the horizontal, vertical, and spherical forms. Zhang et al. studied the crack morphology of the APS-deposited LZ coating using a quantitative imaging analysis method [9]. It was found that the cracks were primarily horizontal in the top and middle regions of the cross-section area, while vertical cracks became dominant in the bottom region. In addition, the calculated porosities showed a nonuniformity in the cross-­ section area. Weber et al. indicated that vertical crack is beneficial in TBC application due to enhanced thermomechanical compliance [10, 11]. The LZ-based TBCs were deposited using the spray pyrolysis method, and the vertical cracks were introduced by decomposing the dried coating layer at 575 °C, as shown in Fig. 1.6d. Moreover, the multilayer coating with vertical cracks was fabricated by the successive deposition and decomposition of multiple thin layers. The heat conduction is slowed in this multilayer coating due to the generated cracks, and the thermal durability can be increased due to the increased thermomechanical compliance [11].

2.2  Mechanical Properties 2.2.1  Elastic Modulus Elastic moduli include Young’s modulus (E), bulk modulus (K), shear modulus (G), and Poisson’s ratio (ν). These elastic moduli can be measured by a depth-sensing micro- and nanoindentation technique or an ultrasound pulse-echo method [6, 12]. In the micro-indentation technique, E can be obtained from the unloading slope by adopting Sneddon’s flat-ended cylindrical punch model [6]. Shimamura et al. studied the moduli for a series of Ln2Zr2O7 material using ultrasound pulse-echo measurement [12]. Shimamura found that the elastic moduli (except for Poisson’s ratio) strongly depend on the atomic radius of rare earth elements for the lanthanide zirconate pyrochlore Ln2Zr2O7. The larger atomic radius corresponds to larger E, K, and G values. La has larger atomic radius than Nd, Sm, and Gd, so LZ has larger elastic modulus than Nd2Zr2O7, Sm2Zr2O7, and Gd2Zr2O7 [12]. Many researchers successfully measured the elastic moduli of the LZ powder and coating, as summarized in Table 2.1. The sample used in Vassen’s work was prepared by pressing the LZ powders at 1350~1400 °C. In this chapter, the Young’s modulus of APS-deposited LZ coating with porosity of 7.53% were measured (detailed discussion in Sect. 2.2.2) [13]. Xu et al. investigated the E of EB-PVD-­ deposited LZ coating [14]. Shimamura et  al. measured the LZ powder using the ultrasound pulse-echo method [12]. Girolamo et al. tested the APS-deposited LZ coating, which was exposed at 1350 °C for 50 h [15].

16

2  Properties of Lanthanum Zirconate–Based Thermal Barrier Coatings

Table 2.1  Mechanical properties of LZ coating versus YSZ coating Properties Young’s modulus

Bulk modulus Shear modulus Poisson’s ratio Vicker’s hardness Microindentation hardness Nanoindentation hardness Fracture toughness

LZ 175 ± 11 GPa (densified powder, Vassen [6]) 156 ± 10 GPa (coating, Zhang [13]) 153 GPa (coating, Xu [14]) 280 GPa (powder, Shimamura [12]) 141 GPa (porous coating, Girolamo [15]) 216 GPa (Shimamura [12]) 109 GPa (Shimamura [12]) 0.28 (Shimamura [12]) 5.51 ± 0.25 GPa (coating, Zhang [13]) 8.83 GPa (coating, Xu [14]) 9.9 ± 0.4 GPa (densified powder, Vassen [6]) 8.8 ± 2.1 GPa (coating, Zhang [13]) 1.1 ± 0.2 MPa m1/2 (densified powder, Vassen [6]) 1.84 MPa m1/2 (coating, Xu [14]) 0.9 MPa m1/2 (densified powder, Jiang [16])

YSZ 210 ± 10 GPa (densified powder, Vassen [6])

– – – – 13 ± 1 GPa (densified powder, Vassen [6]) – 2.0–3.3 MPa·m1/2 (Beshish [17])

The Young’s modulus of the hot-pressed LZ sample is about 15% lower than that of the hot-pressed 8YSZ, which is 210 ± 10 GPa. The low Young’s moduli of LZ are advantageous for reducing thermal stresses, which might compensate CTE mismatch in coatings [6].

2.2.2  Hardness and Fracture Toughness The hardness test can be classified into three categories according to the length scale applied in the measurement: macroindentation hardness, microindentation hardness, and nanoindentation hardness. The measured hardness results of LZ are listed in Table  2.1. The hardness of LZ varies significantly due to the sample density change. The hardness increases with increasing coating density. Fracture toughness is a material property, which describes the ability of fracture resistance to maintain cracks in the material without elongation of cracks. The definition of fracture toughness is that if a sample has a crack in loading, the stress intensity factor increases with increasing load until the unstable crack propagation occurs at a critical value of KI [18]. This critical value is fracture toughness (KIC). The standard measurement method of fracture toughness is the four-point bending test of the bulk sample. Another alternative technique is the indentation method, which is widely used to evaluate the fracture toughness of ceramic and coating

2.2  Mechanical Properties

17

systems [19]. The fracture toughness can be estimated from the indentation test, and the relationship between fracture toughness and hardness can be expressed by the following equation [18]:



K IC  0.16

E P  H c1.5

(2.3)

where E is the elastic modulus, H is the hardness, P is the applied load, and c is the sum of the crack length and one-half of the indenter imprint diagonal. The fracture toughness of LZ powders and coatings summarized in Table 2.1. Hot-pressed 8YSZ has higher hardness and fracture toughness than LZ (microindentation hardness of densified 8YSZ is 13 ± 1, fracture toughness of densified 8YSZ is 2.0–3.3 MPa·m1/2) [6, 17]. The low fracture toughness is the major disadvantage of LZ, which leads to a severe limit on its application as a TBC material. To improve the fracture toughness of LZ material, a composite or multilayer LZ coating can be used. Jiang et al. studied the microstructure and mechanical properties of LZ/Zr0.92Y0.08O1.96 composite ceramics, which were prepared by spark plasma sintering at 1450 °C. The results revealed that the composite ceramics had a higher fracture toughness than single-phase LZ [16]. Jiang et al. also showed that the fracture toughness of composite LZ and 4YSZ coatings increased with increasing the content of 4YSZ. The fracture toughness reached to a value of 1.8 ± 0.1 MPa m1/2 for 50% 4YSZ plus 50% LZ composite coating, which is about two times of that of single-phase LZ coating. In this chapter, we compared single LZ and double-layer LZ/8YSZ coatings [20]. The results showed that the double-layer coating composed of LZ plus porous 8YSZ substantially improved the durability in thermal cycling tests, suggesting the bilayer design is a feasible solution to improve the fracture toughness in LZ-based coatings.

2.2.3  Erosion Resistance Erosion is regarded as a secondary cause of failure of TBCs by deteriorating the coating through progressive removal of the coating material due to the mechanical interaction between the coating surface and the impinging solid particles [21]. The erosion tests by solid particle impingement were standardized by ASTM G76-13 [22]. Typically, the Al2O3 medium with the particle size of 50  μm is used as the abrasive particles, and the abrasive particles are accelerated by the high-velocity carrier gas through a particle-gas supply system. Finally, the abrasive particles are impacted on the surface of the coating at a specific impinging angle. The average erosion rate (mg/g or g/kg) is used to evaluate the erosion resistance performance, which is determined by the slope of TBC mass loss (mg) versus the mass (g) of

18

2  Properties of Lanthanum Zirconate–Based Thermal Barrier Coatings

erodent. Many variables such as velocity, working distance, impact angle, abrasive particles properties, coating hardness, and mechanical properties of coating (H and KIC) affect the erosion results. Ramachandran et al. studied the erosion of APS-deposited YSZ and LZ coatings with different coating porosity, abrasive particles’ velocity, and impact angle [23]. They found that the porosity level is the most predominant factor affecting the erosion rate of the coatings. High porosity in the coating increased erosion rate. The erosion rate increased with the increasing of the abrasive particles’ velocity. One method of improving erosion test performance is microstructure modification. For instance, the columnar microstructure of EB-PVD-deposited coating typically provides the improvement of erosive resistance compared to the “splat” grain microstructure of APS-deposited coating [21]. In addition, the aging of the APS coating decreases the erosion rate and therefore enhances the erosion performance. However, the aging of the EB-PVD coating results in a significant increase of erosion rate due to the sintering of columns [21].

2.3  Thermal Properties 2.3.1  Melting Point Temperature and Specific Heat Capacity The high melting point is an important criterion for TBC material, which is good for the thermal stability during the high-temperature operation environment. Thermal analysis is the commonly used method to detect the melting point, which is performed in sealed tungsten crucibles, and the sample temperature is monitored by a spectro-pyrometer. In addition, high-temperature X-ray diffraction experiments (the Advanced Photon Source, Argonne National Laboratory) can also be used to determine the melting temperature by monitoring the pyrochlore (111) peaks [24, 25]. As shown in Table 2.2, the experimentally measured melting point of the LZ is about 2523–2573 K (2250–2300 °C), which is lower than that of YSZ (2953 K, 2680 °C) [5, 6, 24]. Although the melting point of LZ is lower than that of YSZ, it is still high sufficient for most TBC applications. Differential scanning calorimetry (DSC) is the most widely used technique to precisely measure the specific heat capacity (Cp). Several researchers investigated the specific heat capacity of LZ in different conditions, as shown in Fig. 2.1 [6, 15, 26, 27]. The samples tested in Vassen ‘s work were the densified LZ powder using hot pressing technique at the temperature around 1350~1400  °C.  LZ samples in powder form were used in Bolech and Sedmidubsky’s work. In Girolamo’s work, the specific heat of the LZ coating with a porosity of 11% was measured. The specific heat of 8YSZ is ~ 0.55–0.65 J/g/K in the temperature range of 300–1200 °C, which is larger than that of the LZ [28]. For the TBC materials, small Cp values are preferred to reduce thermal diffusivity.

2.3  Thermal Properties

19

Table 2.2  Thermal properties of LZ coating versus YSZ coating Properties Melting point Specific heat capacity

Thermal conductivity

Coefficients of thermal expansion

LZ 2523–2573 K (Vassen [5, 6], Hong [24]) 0.48 J/g/K (@ 1200 K, Vassen [6]) 0.41 J/g/K (@ 400 K, Bolech [26]) 0.42 J/g/K (@ 400 K, Sedmidubsky [27]) 0.44 J/g/K (@ 1200 K, Girolamo [15]) 1.55 W/m/K (dense, @1273 K, Vassen [6]) 2.15 W/m/K (dense, @ 1273 K, Zhu [7]) 0.68 W/m/K (porosity 11.54%, @ 1173 K, this work, Guo [20]) 0.87 W/m/K (EB-PVD, @ 1273 K, Bobzin [29]) 9.45 × 10−6 /K (273–1473 K, Chen [30]) 9–10 × 10–6 /K (400–1600 K, this work, [20]) 9.0–9.7 × 10–6 /K (400–1600 K, Zhang [31]) 7.6–9.1 × 10–6 /K (400–1400 K, Lehmann [1])

YSZ 2953 K (Vassen [6]) 0.65 J/g/K (@ 1200 K, Khor [28])

2.25 W/m/K (dense @1273 K, Vassen [6]) 0.88 W/m/K (Porous, @ 1173 K, this work, Guo [20]) 1.38 W/m/K (porous, @ 1273 K, Bobzin [29])

11 × 10–6 /K (dense @ 1273 K, Vassen [6]

Fig. 2.1  Specific heat capacity (Cp) of LZ in the temperature range from 0 to 1600 K [6, 15, 26, 27]

20

2  Properties of Lanthanum Zirconate–Based Thermal Barrier Coatings

2.3.2  Thermal Conductivity For TBC materials, thermal conductivity is the most important material property. Thermal conductivity is a measurement of heat flux in a temperature gradient according to Fourier’s law, which provides the fundamental theory for heat conduction analysis [32].





q dT / dx

(2.4)

Later, Debye used the phonon gasses analogy to explain the heat conduction process and derived an expression for thermal conductivity, which shows that thermal conductivity is related to the phonon mean free path (Λ) [33, 34].





cv vs  3

(2.5)

where cv is the specific heat capacity (about constant volume) and vs is the speed of sound. As a result, the thermal conductivity is related to the amount of energy that is carried by a particle, the phonon travel speed, and the distance the phonon travels before scattering [19]. The phonon scattering effect would decrease the thermal conductivity, and the point defect (e.g., oxygen vacancies) would substantially increase the phonon scattering, so the Ln2Zr2O7 pyrochlore structure and doped pyrochlore can have a lower thermal conductivity than that of fluorite structure materials. The most widely used experimental method to measure thermal conductivity is the laser-flash method, which was first proposed by Parker in 1961 [35]. Thermal conductivity can be determined using thermal diffusivity (Dth), specific heat capacity (cp), and density (ρ), which can be independently measured experimentally. During the experiment process, the front surface of the sample is irradiated by a high-energy laser pulse (usually duration time less than 1 ms). The temperature at the back surface of the sample is monitored by an infrared detector [19]. The thermal diffusivity (Dth) can be determined by the bottom surface temperature and time curves using Parker’s theory [35]: Dth 

1.38 L2  2 t0.5

(2.6)

where L is the sample thickness and t0.5 is the time required for the bottom surface to reach half of the maximum temperature rise. The measurement error of the standard flash method is less than 5%. Subsequently, the thermal conductivity (κ) of the LZ coating with given density can be determined according to the following Eq. (2.6):

21

2.3  Thermal Properties



  Dth  c p



(2.7)

Vassen et al. measured the thermal conductivity of LZ samples using the laser-flash method, in which samples were prepared by hot pressing in the temperature range of 1350–1400  °C [1, 6]. The density of the hot-pressed sample was greater than 95%. The thermal conductivity that Vessen measured was ~ 1.5–2.0 W/m/K in the temperature range of 200–1400 °C. Zhu et al. did similar studies for the hot-pressed disk-shape samples using the spray-dried LZ powders [7, 36]. The measured thermal conductivities of the densified LZ were 2–4 W/m/K, which were larger than Vessen’s results, in the temperature range of 200–1400 °C. The thermal conductivity is very sensitive to porosity level. A low porosity leads to a high thermal conductivity. The high thermal conductivity in Zhu’s work may be due to the coating’s lower porosity than that in Vassen’s work. In this chapter, the thermal conductivities of APS-deposited LZ coating with a porosity of 11.54% were measured using the laser-flash method (detailed discussion is in Sect. 2.2.3) [20]. The measured average thermal conductivity of the porous LZ was ~ 0.59–0.68 W/m/K in the temperature range of 297–1172 K (24–899 °C). Chen et al. also measured the thermal conductivity of APS-deposited LZ coating with a lower porosity [30]. Bobzin et al. investigated the thermal conductivity of mixed 7 wt.% YSZ and LZ layers deposited by the EB-PVD method [29]. All of the experimental data are compiled in Fig. 2.2. As shown in Fig. 2.2, the thermal conductivities of the LZ coating decrease with the increase of temperature till 1200 K due to lattice scattering, and the conductivities increase again above 1200 K due to radiation contribution. The conductivities also depend on coating porosity [6, 7, 20, 29, 30]. In Vassen and Zhu’s works, the

Fig. 2.2  Thermal conducitivity of LZ in the temperature range of 273–1700 K [6, 7, 20, 29, 30]

22

2  Properties of Lanthanum Zirconate–Based Thermal Barrier Coatings

porosities of the LZ samples were very low so the thermal conductivities were much larger than those of the porous coatings in this chapter, Bobzin and Chen’s research. In addition to porosity, doped LZ materials were developed to further reduce the thermal conductivity value and improve other thermal and mechanical properties, which enhance the thermal cycling performance. Lanthanide elements are an available option of dopants because they form a similar pyrochlore structure with ZrO2. Lehmann et al. showed that the thermal conductivities of the Nd-, Eu-, Gd-, and Dy-doped LZ (30% of La3+ was substituted) were lower than that of pure LZ below 1000 °C [1]. Consequently, Lehmann indicated that the thermal conductivity was affected by the atomic mass and radius of the substituted and substituting atoms. Bansal et al. suggested that Gd and Yb can replace the La3+ cation to form new pyrochlore structure materials such as La1.7Yb0.3Zr2O7, La1.7Gd0.3Zr2O7, and La1.7Gd0.15Y b0.15Zr2O7 [36]. The Gd- and Yb-doped LZ showed a lower thermal conductivity, a better high-temperature stability up to 1650  °C (3000  °F), and a better sintering resistance than LZ. In addition, Yb, Ce, Y, In, and Sc were reported as dopants for both La3+ and Zr4+ sites [19, 37]. Xiang et al. indicated that (La0.7Yb0.3)2(Zr0.7Ce0.3)2O7 and (La0.2Yb0.8)2(Zr0.7Ce0.3)2O7 had lower thermal conductivities than LZ [37]. Wang found that (La1-x1Yx1)2(Zr1-x2Yx2)2O7, (La1-x1Inx2)2(Zr1-x2Inx2)2O7, and (La1-x1Scx1)2(Zr1-­ x2Scx2)2O7 had the potential to acquire lower thermal conductivity than LZ [19]. The low thermal conductivity is the prime advantage of the LZ coating over the YSZ coating. Vassen et al. compared the thermal conductivity of hot-pressed YSZ and LZ samples [6]. Vassen indicated that high dense LZ samples’ thermal conductivities were 30–35% lower than YSZ at 800–1000 °C, in the similar porosity level. Bobzin et al.’ s work showed that the thermal conductivity of EB-PVD-deposited 7YSZ was about 25–40% larger than the EB-PVD-deposited LZ coating [29]. However, the thermal conductivity of APS-deposited LZ coating was not investigated. In Chaps. 2 and 4, the thermal conductivity of APS-deposited LZ-based TBC coating will be systematically studied using experimental and modeling methods.

2.3.3  Coefficient of Thermal Expansion TBC is a multiple layer coating system including alloy, intermetallic bond coat, TGO layer, and ceramic top coat; therefore, the volume change in the thermal cycling process is different due to the different CTEs in each layer. Thermally induced residual stress generated among TBC layers due to CTE mismatch becomes a primary cause of failure [38]. The CTE values of superalloy substrate and bond coat are usually about 15 × 10−6 and 14 × 10−6 /K, respectively, at 1000 °C [3839], which are much larger than the typical ceramic top coat. As a result, large CTEs are preferred for the ceramic top coat to reduce CTE mismatch. Additionally, due to large thickness of substrate, its CTE influence may be more important than bond coat. The most widely used approach to measure CTEs is the dilatometry method. The linear CTE value can be obtained from the measurements of the temperaturedependent length change of the measured sample using a high-­ temperature

23

2.3  Thermal Properties

dilatometer. The bar-shaped sample is used in the measurement. The specimen is heated at a constant heating rate (normally 5  °C/min), and the temperature and length value are continuously monitored. The linear CTE can be calculated by [40]:



l 

1 dL  L dT

(2.8)

where L is the original length at initial testing temperature, dL represents the unit length change between each scan step, and dT is the unit temperature change between each scan step. Many researchers measured the CTE of LZ using the dilatometer, as summarized in Fig. 2.3 and Table 2.2 [1, 20, 30, 31]. Chen et al. investigated the CTE of both bulk LZ material and APS-deposited LZ coating using the dilatometer, and the results showed a similar trend [30]. The apparent CTE value of the LZ coating was about 9.45 × 10−6 °C−1 in the temperature range of 0~1200 °C. In this chapter, dilatometry method was used to measure the CTE of LZ. The average porosity of the APS-deposited LZ coating in this measurement was 11.54% (detailed discussion is in Sect. 2.2.3) [20]. Zhang et al. determined the CTE value of LZ powders by the variation of lattice parameters during the heating process [31]. The lattice parameters came from the XRD analysis at different temperatures. Lehmann, using a dilatometer, investigated the CTE of hot-pressed LZ samples with density between 69% and 93% of the theoretical value [1].

Fig. 2.3  CTEs of LZ in the temperature range of 273–1600 K [1, 20, 30, 31]

24

2  Properties of Lanthanum Zirconate–Based Thermal Barrier Coatings

The CTE of pure LZ is lower than that of YSZ (11 × 10−6 /K at 1000 °C) [6]. As a result, the CTE difference between the LZ top coat and the bond coat in LZ-based TBC system is larger than that between the YSZ top coat and the bond coat in YSZ TBCs. This is a disadvantage for LZ in TBC application because this may lead to a large volume change during the thermal cycling process, which generates higher thermal and residual stress. However, CTE value of the LZ-based coating can be increased by doping with some rare earth element dopants. For example, Xiang et al. indicated that the CTE of (La0.7Yb0.3)2(Zr0.7Ce0.3)2O7 and (La0.2Yb0.8)2(Zr0.7Ce0.3)2O7 were higher in the high-temperature range (above 400 °C) than that of pure LZ [37]. Meanwhile, Cao and Zhang suggested that using Ce (5–20%) to dope into LZ to increase the CTE value [31, 41].

2.3.4  F  urnace Cycling Test (FCT) and Jet Engine Thermal Shock (JETS) Test Thermal cycling tests are applied to simulate the operation environment of TBCs in gas turbines. Thermal cycling tests can be sorted into two broad categories based on temperature gradient during thermal process [42]. (1) A constant temperature distribution in the sample without a gradient in TBC samples. When heating/cooling rates are low, such as the furnace cycling test (FCT), the sample is slowly heated in a furnace, which creates a high-temperature isothermal environment for the entire TBC system, and then it is cooled by the compressed gas or ambient air cooling out of the furnace [43]. (2) The thermal cycling tests with a temperature gradient across the sample due to fast heating/cooling, such as the jet engine thermal shock (JETS) test, laser rig, and flame rig [42]. In the JETS test, a typical cycle consists of a 20 s heating process, a 20 s forced nitrogen gas cooling, and a 40 s dwell cooling in ambient air environment. The front surface temperature can reach 1400 °C. The failure criterion in the JETS tests is more than 20% spallation of the TBC sample. Since the back side of the sample is not heated, a thermal gradient is generated in the TBC samples during the JETS test. The temperature gradient in the JETS over the whole sample depends on the thickness of the coating system, coating composition, porosity, and microstructure of the coating [43]. Although the maximum temperature and the heating and cooling duration time vary in different thermal cycling tests, large thermal stress and strain mismatch generated due to the CTE mismatch between the top and bond coats are the principal reason for the failure of LZ coatings. To obtain a long thermal cycling lifetime, the LZ coating needs to accommodate the thermal strain associated with thermal cycling [4]. In addition, oxidation of bond coat and low KIC of LZ are additional factors for the spallation of the LZ coating [20, 44]. Many researchers have conducted thermal cycling tests of LZ in different testing conditions. Because the TBC is a complicated system, the thermal cycling results vary due to different cycling test parameters, coating porosities, coating microstructures, multilayer coating architectures, and coating compositions, etc. Vassen et al.

References

25

conducted thermal cycling test with a large temperature gradient across the sample [5, 6, 45]. The APS-deposited LZ coating with single-layer coatings and the APS double-layer coatings with YSZ plus LZ top coat were used in these tests. These TBC systems were tested in the surface temperature range of 1200–1450 °C. The heating and cooling time periods were 5 min and 2 min. The results indicated that the single-layer coating had a rather poor thermal cycling performance. The doublelayer system showed a similar to or slightly better performance than that of YSZ coatings at temperatures below 1300 °C, suggesting the double-layer coating with YSZ is an effective way to improve the lifetime of TBC in thermal cycling tests [45]. Meanwhile, Cao et al. also showed that the single-­layer LZ coating had a short thermal cycling lifetime, but the double-layer LZ coating with La2O3-ZrO2-CeO2 composite sublayer can greatly improve the lifetime [46]. However, the more layers that the TBC coating has, the more artificial defects might be generated during the deposition process. To reduce artificial defects, the composition, thickness, and porosity in the double-layer coatings need to be properly tailored. Bobzin et al. investigated the thermal cycling performance of EB-PVD-deposited LZ and 7YSZ coatings using the FCT test [44]. The samples were heated to 1050 °C for about 30 min, then cooled to 35 °C in compressed air for 5 min. Delamination of the LZ coating occurred at 1856 cycles, which showed a better performance than 7YSZ coating (1380 cycles). The alumina scale was observed in Bobzin’s thermal cycle experiments, which was evidence of bond coat oxidation. Bobzin suggested that the main reason of the failure was a combined effect of oxidation of the bond coat and CTE mismatch. In this chapter, we conducted JETS test for single-layer LZ coating and double-layer coating composed of LZ and 8YSZ deposited by APS with different porosities [20]. The front surfaces of the TBC samples were heated to 1232 °C for 20 s, cooled by compressed N2 gas for 20 s, and followed by ambient cooling for 40 s. The front and back side temperatures were monitored during the test by pyrometers to analyze the cross-sectional temperature gradient. The initial spallation time of the TBC can be pinpointed by the temperature difference between the front and back side surfaces. Our results showed that the single-layer porous 8YSZ coating had better JETS performance than the single-layer porous LZ coating. Additionally, the thermal durability of LZ-based coating can be improved by introducing a porous 8YSZ buffer layer between the top and bond coats [20].

References 1. Lehmann, H., et al. (2003). Thermal conductivity and thermal expansion coefficients of the lanthanum rare-earth-element zirconate system. Journal of the American Ceramic Society, 86(8), 1338–1344. 2. ASTM Standard 328-96. (1996). Standard test method for density, oil content, and interconnected porosity of sintered metal structural parts and oil-impregnated bearings. West Conshohocken: American Society of Testing and Materials. 3. Myoung, S.-W., et  al. (2010). Microstructure design and mechanical properties of thermal barrier coatings with layered top and bond coats. Surface and Coatings Technology, 205(5), 1229–1235.

26

2  Properties of Lanthanum Zirconate–Based Thermal Barrier Coatings

4. Clarke, D., & Levi, C. (2003). Materials design for the next generation thermal barrier coatings. Annual Review of Materials Research, 33(1), 383–417. 5. Vassen, R., et  al. (1999). La2Zr2O7—a new candidate for thermal barrier coatings. In Proceedings of the united thermal spray conference-UTSC’99. Düsseldorf, Germany: DVS-Verlag. 6. Vassen, R., et al. (2000). Zirconates as new materials for thermal barrier coatings. Journal of the American Ceramic Society, 83(8), 2023–2028. 7. Zhu, D., Bansal, N. P., & Miller, R. A. (2003). Thermal conductivity and stability of HfO2-­ Y2O3 and La2Zr2O7 evaluated for 1650 C thermal/environmental barrier coating applications. In Proceedings of the 105th annual meeting and exposition of the American Ceramic Society. Nashville, TN: The American Ceramic Society. 8. Nair, J., et  al. (1999). Sintering of lanthanum zirconate. Journal of the American Ceramic Society, 82(8), 2066–2072. 9. Zhang, J., et al. (2014). Quantitative analysis of pore morphology in lanthanum zirconate thermal barrier coating. Materials Science and Technology (MS&T), 2014, 2061–2068. 10. Weber, S. B., et al. (2013). Lanthanum zirconate thermal barrier coatings deposited by spray pyrolysis. Surface and Coatings Technology, 227, 10–14. 11. Weber, S. B., et al. (2014). Thermal and mechanical properties of crack-designed thick lanthanum zirconate coatings. Journal of the European Ceramic Society, 34(4), 975–984. 12. Shimamura, K., et  al. (2007). Thermophysical properties of rare-earth-stabilized zirconia and zirconate pyrochlores as surrogates for actinide-doped zirconia. International Journal of Thermophysics, 28(3), 1074–1084. 13. Zhang, J., et al. (2014). Microstructural non-uniformity and mechanical property of air plasma-­ sprayed dense lanthanum zirconate thermal barrier coating. Materials Today: Proceedings, 1(1), 11–16. 14. Xu, Z., et al. (2010). Preparation and characterization of La2Zr2O7 coating with the addition of Y2O3 by EB-PVD. Journal of Alloys and Compounds, 492(1–2), 701–705. 15. Di Girolamo, G., et  al. (2015). Evolution of microstructural and mechanical properties of lanthanum zirconate thermal barrier coatings at high temperature. Surface and Coatings Technology, 268, 298–302. 16. Jiang, K., et al. (2014). Microstructure and mechanical properties of La2Zr2O7–(Zr0.92Y0.08) O1.96 composite ceramics prepared by spark plasma sintering. Ceramics International, 40(9, Part A), 13979–13985. 17. Beshish, G. K., et al. (1993). Fracture toughness of thermal spray ceramic coatings determined by the indentation technique. Journal of Thermal Spray Technology, 2(1), 35–38. 18. Anstis, G., et al. (1981). A critical evaluation of indentation techniques for measuring fracture toughness: I, direct crack measurements. Journal of the American Ceramic Society, 64(9), 533–538. 19. Wang, Y. (2013). The improvement of thermal and mechanical properties of La2Zr2O7-based pyrochlores as high temperature thermal barrier coatings (PhD thesis). School of Materials, The University of Manchester: Manchester, UK. 20. Guo, X., et  al. (2016). Thermal properties, thermal shock, and thermal cycling behavior of lanthanum zirconate-based thermal barrier coatings. Metallurgical and Materials Transactions E, 3(2), 64–70. 21. Wellman, R., & Nicholls, J. (2007). A review of the erosion of thermal barrier coatings. Journal of Physics D: Applied Physics, 40(16), R293. 22. ASTM G76-13. (2015). Standard test method for conducting erosion tests by solid particle impingement using gas jets. West Conshohocken: American Society of Testing and Materials. 23. Ramachandran, C. S., Balasubramanian, V., & Ananthapadmanabhan, P. V. (2013). Erosion of atmospheric plasma sprayed rare earth oxide coatings under air suspended corundum particles. Ceramics International, 39(1), 649–672. 24. Hong, Q.-J. (2015). Methods for melting temperature calculation. Pasadena: California Institute of Technology.

References

27

25. Ushakov, S. V., & Navrotsky, A. (2012). Experimental approaches to the thermodynamics of ceramics above 1500 C. Journal of the American Ceramic Society, 95(5), 1463–1482. 26. Bolech, M., et al. (1997). The heat capacity and derived thermodynamic functions of La2Zr2O7 and Ce2Zr2O7 from 4 to 1000 K. Journal of Physics and Chemistry of Solids, 58(3), 433–439. 27. Sedmidubsky, D., Benes, O., & Konings, R.  J. M. (2005). High temperature heat capacity of Nd2Zr2O7 and La2Zr2O7 pyrochlores. The Journal of Chemical Thermodynamics, 37, 1098–1103. 28. Khor, K. A., & Gu, Y. W. (2000). Thermal properties of plasma-sprayed functionally graded thermal barrier coatings. Thin Solid Films, 372(1–2), 104–113. 29. Bobzin, K., et al. (2013). Influence of temperature on phase stability and thermal conductivity of single- and double-ceramic-layer EB–PVD TBC top coats consisting of 7YSZ, Gd2Zr2O7 and La2Zr2O7. Surface and Coatings Technology, 237, 56–64. 30. Chen, H., et al. (2009). Thermophysical properties of lanthanum zirconate coating prepared by plasma spraying and the influence of post-annealing. Journal of Alloys and Compounds, 486(1–2), 391–399. 31. Zhang, J., et al. (2012). Thermal expansion and solubility limits of cerium-doped lanthanum zirconates. Journal of Alloys and Compounds, 525, 78–81. 32. Bergman, T. L., et al. (2011). Fundamentals of heat and mass transfer. Hoboken: Wiley. 33. Debye, P.  J. W., et  al. (1914). Vorträge über die kinetische Theorie der Materie und der Elektrizität (Vol. 6). BG Teubner. 34. Kittel, C. (1949). Interpretation of the thermal conductivity of glasses. Physical Review, 75(6), 972–974. 35. Parker, W. J., et al. (1961). Flash method of determining thermal diffusivity, heat capacity, and thermal conductivity. Journal of Applied Physics, 32(9), 1679–1684. 36. Bansal, N.  P., & Zhu, D. (2007). Effects of doping on thermal conductivity of pyrochlore oxides for advanced thermal barrier coatings. Materials Science and Engineering: A, 459(1–2), 192–195. 37. Xiang, J., et al. (2012). Phase structure and thermophysical properties of co-doped La2Zr2O7 ceramics for thermal barrier coatings. Ceramics International, 38(5), 3607–3612. 38. Padture, N.  P., Gell, M., & Jordan, E.  H. (2002). Thermal barrier coatings for gas-turbine engine applications. Science, 296(5566), 280–284. 39. Taylor, T.  A.. (2011). Low thermal expansion bondcoats for thermal barrier coatings. US Patents No. 7910225 B2. 40. Kutty, K. V. G., et al. (1994). Thermal expansion behaviour of some rare earth oxide pyrochlores. Materials Research Bulletin, 29(7), 759–766. 41. Cao, X. Q., Vassen, R., & Stoever, D. (2004). Ceramic materials for thermal barrier coatings. Journal of the European Ceramic Society, 24(1), 1–10. 42. Vaßen, R., et  al. (2012). Testing and evaluation of thermal-barrier coatings. MRS Bulletin, 37(10), 911–916. 43. Bolcavage, A., et al. (2004). Thermal shock testing of thermal barrier coating/bondcoat systems. Journal of Materials Engineering and Performance, 13(4), 389–397. 44. Bobzin, K., Lugscheider, E., & Bagcivan, N. (2006). Thermal cycling behaviour of lanthanum zirconate as EB-PVD thermal barrier coating. Advanced Engineering Materials, 8(7), 653–657. 45. Vaßen, R., Traeger, F., & Stöver, D. (2004). New thermal barrier coatings based on pyrochlore/YSZ double-layer systems. International Journal of Applied Ceramic Technology, 1(4), 351–361. 46. Wang, L., et  al. (2012). Thermal shock behavior of 8YSZ and double-ceramic-layer La2Zr2O7/8YSZ thermal barrier coatings fabricated by atmospheric plasma spraying. Ceramics International, 38(5), 3595–3606.

Chapter 3

Modeling of Mechanical Properties of Lanthanum Zirconate Crystal

3.1  Motivation and Introduction Although the low thermal conductivity of LZ is an advantage for the application of TBC, LZ was also reported that it have lower Young’s modulus, hardness, and lower fracture toughness than 8YSZ [1]. As discussed in Chap. 2, the thermal cycling performance of the single-layer LZ TBC are worse than that of 8YSZ, which might due to the low coefficient of thermal expansion (CTE) and poor mechanical properties of LZ [2, 3]. As a result, the investigation of the mechanical property of LZ would be very important for systematically understanding the LZ materials. Tensile and shear stress-strain relations of single-crystal LZ exhibit many anisotropic mechanical properties. However, single-crystal LZ samples are hard to be obtained. Since it is hard to measure the intrinsic mechanical properties of perfect LZ in experiment due to its porous character, theoretical modeling techniques can be used to study these mechanical properties. The most widely used theoretical method to investigate the mechanical ­properties is the analytical method based on Voigt-Reuss-Hill’s theories, which calculates the elastic moduli from the elastic constants [4–6]. The elastic constants can be computed using the first-principle calculations based on density functional theory (DFT) [7, 8]. In addition, the nanoscale tensile and shear simulation can be conducted using DFT and MD calculations directly. The uniaxial nanoscale-tensile strain can be applied stepwise in the tensile direction. The shear strain can be applied stepwise by changing the crystal angle. After the full relaxation at each strain stage, the tensile and shear stresses can be computed using the DFT or MD calculations. Although several methods were developed to conduct the nanoscale tensile and shear simulation, to our best knowledge, no such investigation was performed on LZ material. In this book, the single LZ crystal model was built and the lattice parameter of LZ conventional unit cell was calculated by minimizing the total system energy. The nanoscale tensile and shear stress-strain relations of single-crystal LZ were investigated using DFT and MD method. The corresponding elastic moduli © Springer Nature Switzerland AG 2021 X. Guo et al., Novel Lanthanum Zirconate-based Thermal Barrier Coatings for Energy Applications, https://doi.org/10.1007/978-3-030-58695-9_3

29

30

3  Modeling of Mechanical Properties of Lanthanum Zirconate Crystal

were calculated using two approaches: stress-strain curve analysis in large deformation and Voigt-Reuss-Hill approximation in small deformation [4, 9]. ­ Toughness was calculated based on the stress and strain curves as well. Average Bader charge difference and charge density distribution were used to explain the influence of electron interaction on the Young’s modulus.

3.2  Theoretical LZ Model 3.2.1  DFT Model In this book, the DFT calculations were conducted using Vienna Ab initio Simulation Program (VASP) [10, 11]. The exchange-correlation potential was specified using projector augmented wave (PAW) method of generalized gradient approximation (GGA) functional. The Brillouin zone k-point of was conducted using 3 × 3 × 3 Monkhorst-Pack scheme. A conjugate-gradient algorithm was used for the ironic relaxation. The plan-wave cutoff energy was 500 eV. The energy relaxation criterion for the electron was 10−6 eV, and the ionic relaxation convergence criterion was total forces smaller than 0.01 eV/A. The LZ crystal is a cubic pyrochlore structure, which belongs to space group of Fd 3 m [12]. There are four independent crystallographic atom sites, where La is at 1 1 1 1 1 16d at ( , , ), Zr is at 16c at (0,0,0), O1 is at 48f at (x, , ), and O2 is at 8b 2 2 2 8 8 3 3 3 at ( , , ). The x value of O1 varies from 0.3125 to 0.375. According to Tabira’s 8 8 8 experiments, 0.333 was chosen as the initial x value of LZ in this book [13]. The conventional LZ unit cell was used in the elastic constant calculation and the shear stress-strain calculations, as shown in Fig. 3.1. Fig. 3.1  Crystal structure of LZ unit cell. Each ball shown as green, red, and blue indicates Zr atom, O atom, and La atom, respectively (same coloring schemes are used in all figures afterward)

3.2  Theoretical LZ Model

31

To calculate the stress in a specific crystal direction, the LZ unit cell was rebuilt to align the (001), (011), and (111) surface perpendicular to the tensile axis for the tensile stress-strain calculations, as shown in Fig.  3.2. The atom numbers of the (001), (011), and (111) model is 44, 44, and 66, respectively. All of the LZ structures were relaxed before the elastic constant, tensile, and shear calculations. Large deformation stress-strain analysis and the small deformation elastic constant analysis were used to calculate the anisotropic elastic moduli. In the stress-­ strain analysis, the anisotropic Young’s modulus and shear modulus in a particular tensile and shear direction were obtained from the slope of the linear fitted line in the elastic stage of the stress-strain curves. In the small deformation elastic constant analysis, Voigt-Reuss-Hill approximation were used to calculate the elastic moduli of LZ base on the elastic constants of single LZ unit cell. The elastic constants of LZ unit cell were calculated using DFT method by performing finite distortion of the LZ unit cell [14]. The Voigt-proposed approach expresses the stress in a single crystal in terms of the given strain. On the other hand, the Reuss-proposed approach expresses the strain in terms of the given stress. The Voigt-Reuss-Hill approximation calculated the effective moduli of an aggregated polycrystalline combined the Voigt and Reuss’s approximation [4, 15]. For cubic single-crystal LZ, only three

Fig. 3.2  Tensile model at (a) [001] direction along a axis, (b) [011] direction along c axis, and (c) [111] direction along c axis directions. Each ball shown as green, red, and blue indicates Zr atom, O atom, and La atom, respectively

32

3  Modeling of Mechanical Properties of Lanthanum Zirconate Crystal

elastic constants are independent (C11, C12, C44). The effective bulk modulus (K), shear modulus (G), Young’s modulus (E), and Poisson’s ratio (υ) of cubic LZ crystal can be calculated using following equations:



C11  2C22 3

(3.1)

G

GR  GV 2

(3.2)

E

9 KG 3K  G

(3.3)

K





3 K  2G 2  3K  G 

(3.4)

where GV and GR are the shear modulus expressed by the Voigt and Reuss approach, respectively [4, 15].





GV 

C11  C22   3C44 5

5 4 3   GR C11  C22 C44



(3.5) (3.6)



The slip system (easy slip plane and direction) in the shearing process often occurs on the plane of high atomic density in closely packed directions [16]. For example, the primary slip system for face center cubic (fcc) crystal is {111} and the secondary slip direction is {111}. Because the {111} is the closest packed plane in LZ crystal, the most possible slip direction is along {111} and {111} directions. The conventional unit cell of LZ was used in the shear stress calculation, as shown in Fig.  3.1. The shear directions were controlled by rotating the LZ unit cell to align the slip plane perpendicular to one of the coordination axis. The shear strain was applied by changing the angle in the shear directions. The shear stress was calculated using first-principle calculations corresponding to each strain step. Bader charge transfer analysis in tensile and shear DFT model was conducted to describe the electron gain and loss between different atoms [17, 18]. The average charge differences between Zr, La, and O were calculated from the Bader charge results, which gave the insights of atom bonding character in different orientation.

3.2  Theoretical LZ Model

33

3.2.2  MD Model However, all of the DFT calculations for tensile and shear are performed on ground state of atoms, which is typically at temperature of 0 K. The tensile and shear simulations at a higher temperature were conducted using molecular dynamics (MD) calculation, and MD also can be performed at a larger scale than the first-principle calculations. All of the MD calculations in this study were conducted using LAMMPS (Large-scale Atomic/Molecular Massively Parallel Simulator) MD simulation package [19]. In this work, the interaction between atoms was assumed to be purely ionic, so the Buckingham and Coulomb potential was used to describe the short- and long-range atom interactions, respectively, which can be expressed as follows:  rij Uij  rij   Aij exp    ij



2  Cij 1 qi q j e    6  rij  rij 4  0

(3.7)

where rij is the distance of the interactive ions; qi and qj are the charges of La3+, Zr4+, O2− ions; ρ is an ionic-pair-dependent length parameter; and A and C are adjustable parameters. The parameters of the Buckingham potential used in this study are listed in Table 3.1, which were fitted based on experimental crystallographic data and Hartree-Fock calculated elastic constants by Crocombette et al. [20]. Both tensile and shear simulations were conducted in 300  K.  The tensile simulation was carried out using uniaxial elongation along x axis, and the shear simulation was performed using uniaxial compression. Periodical boundary condition was used in all these MD simulations. The first-principle calculation of LZ unit cells shown in Fig. 3.2 was used to assemble the MD tensile and shear models. For tensile MD calculation, three tetragonal boxes of 12 × 12 × 24 (individually in x, y, z directions) LZ unit cells in [001, 110, 111] directions were assembled as tensile models in each direction. The MD model of tensile simulation in [001] direction is shown in Fig. 3.9a. In the tensile model, the (001), (110), and (111) surfaces were perpendicular to the tensile axis (z). The shear models were built with 16 × 16 × 16 LZ unit cells in [112] direction, and 24 × 36 × 16 LZ unit cell in [110] direction. The compression axis was performed along [110] and [11 2 ] directions. The conjugate gradient method was used to minimize the energy of the system in both tensile and shear simulations [21]. Then, the time integration was performed in isothermal-­isobaric (NPT) ensembles, which was dedicated to generate the position and v­ elocities parameters. Finally, the tensile models were Table 3.1  Buckingham potential parameter for LZ [140] Interactions O-O La-O Zr-O

A (eV) 22764.00 1367.41 1478.69

Ρ (Å) 0.1490 0.3591 0.3554

C (eV∙ Å6) 27.89 0.00 0.00

Charges −2.0 −3.0 +4.0

34

3  Modeling of Mechanical Properties of Lanthanum Zirconate Crystal

elongated in the NPT ­ensembles, and the shear models were compressed in the canonical (NVT) ensembles. The time step of all the MD simulations performed at 1 femtosecond.

3.3  LZ Crystal Constant Optimization The crystal constant or geometry optimization was calculated by minimization of the total system energy. The interatomic potential energy is the sum of the repulsive energy and attractive energy. The potential energy (E) as a function of interatomic distance (r) is shown in Fig. 3.3 [22]. In a two-atoms system, when r is smaller than the equilibrium distance r0, the repulsive energy increases and attractive energy decreases which results in a repulsive force on the atoms. Therefore, the atom pair will be pushed away. On the other hand, when r becomes larger than r0, the attractive energy takes advantage to pull the atoms together. When the equilibrium distance r0 between the two atoms is reached, the total energy reaches its lowest value. In a crystal system, the crystal constant that has the minimum total energy can be regarded as the equilibrium state of the system, which means the system has the highest possibility to balance the repulsive and attractive forces among atoms. As a result, the lattice constants can be obtained by analysis relations between the total energy and the lattice contents. Since LZ is a cubic crystal, it has only one lattice constant. The actual LZ lattice constant was determined by finding the minimum point on the curve of the total energy as a function of lattice constant. Figure  3.4 shows DFT-calculated total energy of single LZ conventional cell as a function of lattice constant. This work tested the lattice constant of LZ model from 9 Å to 12 Å, which has a wide enough range to contain the actual LZ lattice constant. The calculation results indicated that the equilibrium state of the system occurred at the lattice constant of 10.875  Å, where the total energy reached its minimum point of −803.749 eV. Fig. 3.3  Potential energy as a function of interatomic distance

35

3.3  LZ Crystal Constant Optimization

Fig. 3.4  Total energy as a function of LZ lattice constant Table 3.2  Analytically calculated lattice constant from XRD peaks XRD peak Lattice constant (Å)

(2 2 2) 10.814

(4 0 0) 10.799

(4 4 0) 10.802

(6 2 2) 10.798

(4 4 4) 10.786

(8 0 0) 10.804

To validate the DFT-calculated results, the lattice constant of LZ powders was analytically calculated from XRD experimental data, as shown in Fig. 2.2 (Sect. 2.1.2 LZ Powder characterizations). Bragg’s law was used to analyze the XRD data: n  2 dsin



(3.8)

where λ = 1.54 Å for the laboratory XRD used in this work, n = 1. The lattice parameters can be derived by d and Miller indices of each XRD peak. Specifically, for cubic unit cell, lattice parameter (a0) can be expressed by:



dhkl 

a0 , where s  h 2  k 2  l 2 s

(3.9)

Table 3.2 shows the calculated lattice constant from several XRD peaks. The average lattice constant of LZ powders obtained from XRD experimental data is 10.801 Å. Shimamura and Tabira’s experiment results of the LZ lattice constant are 10.8 Å and 10.802 Å, respectively, which are the same as the XRD experiment results in this work [13, 23]. The difference between the DFT-calculated result and the XRD experiment result is 0.68%. The DFT-calculated lattice constant of LZ is very accurate.

36

3  Modeling of Mechanical Properties of Lanthanum Zirconate Crystal

3.4  Stress-Strain Analysis and Anisotropic Elastic Moduli 3.4.1  Large Deformation Stress-Strain Analysis As mentioned in Sect. 3.2, the single LZ cell nanoscale models used for DFT tensile and shear simulations were (001), (110), (111), and conventional LZ cell, respectively. The ions positions of all the DFT models were relaxed using conjugate-­gradient algorithm before and in the deformation calculations. In the tensile simulation, the model is uniaxial elongated in z direction, so the tensile stress acquired directly from stress tensor in z direction. Figure 3.5 shows the tensile stress-strain curves calculated using first-principle calculations. The linear elastic stage on these curves only occurred in the first a few steps, then the stress drops slowly after it reached the ultimate tensile strength. The LZ tensile model in [001] direction has the largest ultimate tensile strength, and the LZ tensile models in [110, 111] directions have similar ultimate strength. Young’s modulus can be calculated form the slope of linear section in the curve using Hooke’s law. The Young’s modulus in [001, 110, 111] directions were obtained from the slope of the linear fitted line, which corresponds to the strain from 0 to 0.1 for [001] direction, and from 0 to 0.08 for [110, 111] directions. The toughness can be calculated from the integration of the area under the stress-strain curve. The calculated Young’s modulus, ultimate tensile strength, and toughness are summarized in Table  3.3. The single-crystal LZ has the maximum tensile Young’s modulus in [111] direction, and minimum Young’s modulus in [001] plane, this trend is consistent with Liu’s results [24]. In the nanoscale DFT simulations results, the quasi-plastic deformation stage occurred on the stress-strain curves rather than the ideally

Fig. 3.5  DFT-calculated tensile stress-stain curve

37

3.4  Stress-Strain Analysis and Anisotropic Elastic Moduli Table 3.3  Calculated elastic modulus, ultimate strength, and toughness Calculation model DFT Tensile [001] DFT Tensile [110] DFT Tensile [111] MD Tensile [001] MD Tensile [110] MD Tensile [111] DFT Shear {111} DFT Shear {111} MD Shear {111} MD Shear {111}

Elastic modulus (GPa) 207.211 213.225 222.596 230.355 226.897 229.751 76.037 73.381

Ultimate strength (GPa) 47.747 23.453 20.980 30.250 18.867 18.914 22.507 20.192

Toughness (MJ/m3) 21.565 12.219 8.934 3.018 2.192 1.839 8.862 8.028

– –

29.008 20.774

– –

e­ lastic-brittle type [25]. The stress and strain were largely increased during the quasi-plasticity stage, which determined the magnitude of ultimate strength and toughness. Although the Young’s modulus in [001] direction was lower than that in [110, 111] directions, the tensile stress-strain curve in [001] direction possessed a higher “yield” point in quasi-­plasticity stage than that in [110, 111] directions. As a result, the LZ tensile model in [001] direction possessed the maximum ultimate strength and toughness. The ultimate strength and toughness are properties of large deformation; however, the Young’s modulus is a property of small elastic deformation. These properties describe stress-strain relations under different deformation circumstances, so they can exhibit different trend. Figure 3.6 shows the DFT-calculated shear stress-strain curves in two slip system {111} and {111}. Basically, the two curves are very similar. As summarized in Table 3.3, the shear modulus of {111} is slightly larger than that of {111}, and so is the ultimate shear strength and the toughness, although the differences are very small. When the two slip systems of LZ crystal are under same stress level, the corresponding strain in {111} direction will be larger than that in {111} direction. Based on these DFT-calculated results, in single-­ crystal LZ material, the {111} direction system is more likely to slip in the shear simulation.

3.4.2  Small Deformation Elastic Moduli Calculation The elastic constants C11, C12, and C44 of single-crystal LZ were calculated using DFT method in small deformation condition, which were 256.001 GPa, 116.425 GPa, and 86.708 GPa, respectively. The effective elastic moduli of LZ were calculated using Eq. 3.1–3.6, as summarized in Table 3.4. Selected experiments results from literatures were also listed in Table 3.4 as a comparison. In these literature results, Zhang et  al. measured the Young’s modulus of 7% porous LZ coating using

38

3  Modeling of Mechanical Properties of Lanthanum Zirconate Crystal

Fig. 3.6  DFT-calculated shear stress-stain curve Table 3.4  Effective moduli of LZ and the experimental results Young’s modulus (GPa)

Effective moduli 205.12

Bulk modulus (GPa) Shear modulus (GPa) Poisson’s ratio

162.95 79.49 0.29

Experimental results 156 ± 10 (Zhang [26]) 280 (Shimamura [23]) 216 (Shimamura [23]) 109 (Shimamura [23]) 0.28 (Shimamura [23])

­nanoindentator [26], and Shimamura et al. studied the moduli of LZ materials using the ultrasound pulse-echo measurement [23]. Comparing with the experimental data from the literatures, the effective moduli of the LZ are in a reasonable range. The anisotropic characteristics can be determined by Zener anisotropic ratio, which is defined as Z = 2C44/(C11–C22) [5]. If the Z = 1, the material is isotropic. The calculated Z value for LZ is 1.24, indicating that the elastic moduli of LZ is anisotropic, and the maximum Young’s modulus is in direction. Anisotropic Young’s modulus of cubic LZ crystal can be expressed using the DFT-calculated elastic constants, which is given by the following equations [5].







2 2 2 2 2 2 1 l 4  m 4  n4 2 m n  n l  l m   E E0 F0

C11  C22 1  E0  C11  C22   C11  2C22 



(3.10) (3.11)



39

3.4  Stress-Strain Analysis and Anisotropic Elastic Moduli



2C12 2 1   F0  C11  C22   C11  2C22  C44

(3.12)

where l, m, n are directional cosines. The (110) plane was chosen in this work for the anisotropic calculation, because it includes all three of the principal directions in a cubic crystal , , and . The directional cosines for (110) plane can be calculated by: l  cos  m 

sin  2

n

sin  2

(3.13)

where θ is the angle in (110) plane, which measured from [100] direction to the [110] direction. The anisotropic Young’s modulus results were plotted in Fig. 3.7, including the solid curve calculated from the elastic constants of single-crystal LZ in small

Fig. 3.7  Summary of anisotropic Young’s modulus

40

3  Modeling of Mechanical Properties of Lanthanum Zirconate Crystal

d­ eformation status, and the scattered points calculated from the DFT and the MD (MD results will be discussed in Sect. 3.4.4) stress-strain curves in large deformation tensile simulation. As shown in Fig. 3.7, most of anisotropic Young’s modulus results are very close to each other, suggesting that this modeling study has good consistency and accuracy.

3.4.3  Bader Charge Analysis and Charge Density Distribution Bader charge transfer was calculated according to each atom in the tensile and shear systems. The unit of Bader charge is electron (e), which means that the number of the electrons one atom gains or losses. The average Bader charge difference between cation and anion reveals the electron interaction between those two type of atoms, which is related with the bonding character of those atoms. The high average Bader charge difference indicates the strong bond. Table 3.5 summarized the Bader charge differences between cation and anion results in each tensile and shear calculations. The charge differences between Zr and O, and between La and O in tensile [111] calculation are the largest among the three tensile cases, suggesting the bonding strength in tensile model in [111] direction are the strongest. The strong bond between the cation and the anion in the [111] tensile model corresponding to the high Young’s modulus results in DFT calculations. In the shear models, the average Bader charge difference in LZ shearing model in {111} direction is slightly larger than that in {111} direction, suggesting the atom bond in {111} shear model is stronger than that in {111} direction. The stronger bond of shear model in {111} direction results to higher Young’s modulus than that of shear model in {111} direction. The contour of charge density distribution in each LZ tensile simulation with strain of 0.5 are plotted in Fig. 3.8. Charge density distribution reveals the electron loss and gain property. The red color indicates electron gain, and the blue color indicates the electron loss. As shown in Fig. 3.8, the Zr atoms lost almost all its valence electrons to the O atoms. The density distribution near O atoms are similar in the [001] tensile model and [110] tensile model. However, the [111] tensile model Table 3.5  Calculated average Bader charge difference between O, Zr, and La atoms

Tensile in [001] Tensile in [110] Tensile in [111] Shear in {111} Shear in {111}

Young’s modulus (DFT stress-strain analysis) 188.139 189.556 208.462 76.037

Average Bader charge difference between O and Zr 4.592 4.804 5.106 4.965

Average Bader charge difference between O and La 3.469 3.603 3.696 3.568

73.381

4.893

3.535

3.4  Stress-Strain Analysis and Anisotropic Elastic Moduli

41

Fig. 3.8  Charge density distribution (e/Å3) of tensile calculation at strain of 0.5  in (a) [001] ­direction, (b) [110] direction, and (c) [111] direction

has the highest charge density near O atom among all the tensile models, suggesting the electron interaction in [111] tensile model are the strongest. This strong electron interaction leads to a high Young’s modulus in [111] direction.

3.4.4  MD Tensile and Shear Simulations Figure 3.9 shows the models of the first and last step of the MD tensile calculations in [001] direction. The model image of initial step demonstrates the perfect singlecrystal LZ without any tensile strain. The model image of 0.16 strain step shows that there are some vastly deformed areas in the model, which reach its critical strength and are about to break apart.

42

3  Modeling of Mechanical Properties of Lanthanum Zirconate Crystal

Fig. 3.9  MD-calculated tensile model in [001] direction (a) the initial step (no strain), and (b) the step with 0.16 strain. Each ball shown as green, red, and blue indicates Zr atom, O atom, and La atom, respectively

Fig. 3.10  The MD-calculated tensile stress-strain curves

Figure 3.10 shows the MD -calculated tensile stress and strain curves in [001, 011, 111] directions. All of the stress-strain curves in these three tensile simulations

3.5 Summary

43

exhibit an elastic stage and the stress drops after the maximum tensile stress is reached. The slopes of these stress-strain curves in the elastic stage are very similar, when the strain is less than 0.03 at the beginning. Then, the slope changes as the increase of the stress, when the deformation occurs in the tensile model. The Young’s modulus of LZ in [001, 011, 111] directions is calculated from the slope of elastic stage on the MD-calculated stress-strain curves, whose strains ranges from 0 to 0.03, as summarized in Table 3.3. The Young’s modulus values in [001, 111] directions are very close to each other, which are higher than that in [110] direction. The stress-strain curve in [001] direction shows longer linear elastic stage and higher ultimate stress than the stress-strain curves in other two directions, indicating that LZ has a larger toughness in [001] direction than that in [001, 110] directions. The MD-calculated ultimate tensile strengths share the same trend with the DFT results. However, the magnitude of Young’s modulus, ultimate strength, and toughness results are different with DFT results, as listed in Table 3.3. This is primarily due to the different scale in dimension and atom number between DFT and MD calculations. The MD models used in this work are about several hundreds times larger than the DFT model in both atom number and dimensions. The shearing stress was derived from the virial theory by averaging the virial stress over the whole system, which include both the potential and kinetic energies [27]. The shear stress was calculated by following equation:

  0.5  xx  0.5  yy   zz  



(3.14)

where x is the uniaxial compression axis [28]. The relation between the resolved compression and shear slip system was also considered. The orientation of the slip plane in the LZ normal compression model is very close to the initial compression axis, so the compressive shear strain was used to specify the strain in the slip plane as a simplification [28]. Figure 3.11 shows the MD-calculated shear stress-strain curves in [110, 112] directions. The shear stress-strain curves have similar form as the tensile stress-­ strain curve, which show a linear elastic stage. The ultimate shear strength of [110] are obviously larger than that of [112]. Similar as the DFT-calculated shear stress-­ strain curves, the MD shear results also indicate that single-crystal LZ in [112] direction is more likely to slip than in [110] direction.

3.5  Summary In this chapter, the conventional unit cell of single-crystal LZ was built, and the lattice constant of LZ unit cell was optimized using DFT calculations. Based on the LZ unit cell model, the nanoscale tensile and shear simulation for single-crystal LZ material were performed using DFT and MD calculations, and the corresponding mechanical properties were calculated using two approaches: stress-strain curve

44

3  Modeling of Mechanical Properties of Lanthanum Zirconate Crystal

Fig. 3.11  The MD-calculated shear stress-stain curves of LZ in [110, 112] directions

analysis in large deformation and Voigt-Reuss-Hill approximation in small deformation. The DFT-calculated results of both methods are very close, suggesting the consistency of these two methods. Comparing with experimental results from the literature, the effective elastic moduli calculated from Voigt-Reuss-Hill approximation have a good accuracy. The main conclusions of this chapter can be summarized as following aspects. 1. The actual lattice constant of LZ was calculated by minimization of the total energy. The DFT-calculated LZ lattice constant is 10.875 Å, which has an error of 0.68% comparing to the experimental result. 2. The elastic moduli of single-crystal LZ are anisotropic. The DFT results show that the largest Young’s modulus occurred in [111] direction. The average Bader charge difference analysis indicates that the bonds between Zr and O, and between La and O are the strongest in LZ tensile model in [111] direction. The charge density distribution analysis exhibits the strong electron interaction in tensile [111] model, which leads to a high Young’s modulus. 3. The DFT-calculated shear stress-strains are similar in both {111} and {111} slip systems. This is because the average Bader charge differences between the cations and anions in these two systems are similar. The most likely slip system in {111} plane is {111}, because more strain is generated in {111} shear model than that in {111} shear model under the same stress level. 4. The MD-calculated tensile and shear stress-strain curves show the same trend as the DFT-calculated results. However, the value of the mechanical properties calculated using MD is quite different from the DFT results, because the scale in dimensions and number of atoms are different between the MD model and the DFT model.

References

45

References 1. Vassen, R., et al. (2000). Zirconates as new materials for thermal barrier coatings. Journal of the American Ceramic Society, 83(8), 2023–2028. 2. Guo, X., et al. (2016). Thermal properties, thermal shock, and thermal cycling behavior of lanthanum zirconate-based thermal barrier coatings. Metallurgical and Materials Transactions E, 3(2), 64–70. 3. Vaßen, R., Traeger, F., & Stöver, D. (2004). New thermal barrier coatings based on pyrochlore/YSZ double-layer systems. International Journal of Applied Ceramic Technology, 1(4), 351–361. 4. Hill, R. (1952). The elastic behaviour of a crystalline aggregate. Proceedings of the Physical Society. Section A, 65(5), 349. 5. Ingel, R. P., & Iii, D. L. (1988). Elastic anisotropy in zirconia single crystals. Journal of the American Ceramic Society, 71(4), 265–271. 6. Anderson, O.  L. (1963). A simplified method for calculating the Debye temperature from elastic constants. Journal of Physics and Chemistry of Solids, 24(7), 909–917. 7. Hohenberg, P., & Kohn, W. (1964). Inhomogeneous electron gas. Physical Review, 136(3B), B864. 8. Kohn, W., & Sham, L. J. (1965). Self-consistent equations including exchange and correlation effects. Physical Review, 140(4A), A1133–A1138. 9. Wu, L., & Zhang, J. (2015). Ab initio study of anisotropic mechanical properties of LiCoO2 during lithium intercalation and deintercalation process. Journal of Applied Physics, 118(22), 225101. 10. Kresse, G., & Furthmüller, J. (1996). Efficiency of ab-initio total energy calculations for metals and semiconductors using a plane-wave basis set. Computational Materials Science, 6(1), 15–50. 11. Kresse, G., & Furthmüller, J. (1996). Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set. Physical Review B, 54(16), 11169–11186. 12. Moriga, T., et  al. (1989). Crystal structure analyses of the pyrochlore and fluorite-type Zr2Gd2O7 and anti-phase domain structure. Solid State Ionics, 31(4), 319–328. 13. Tabira, Y., et al. (2000). Systematic structural change in selected rare earth oxide pyrochlores as determined by wide-angle CBED and a comparison with the results of atomistic computer simulation. Journal of Solid State Chemistry, 153(1), 16–25. 14. Kresse, G., Marsman, M., & Furthmüller, J. (2013). VASP the guide. http://cms.mpi.univie. ac.at/vasp/vasp/ 15. Den Toonder, J., Van Dommelen, J., & Baaijens, F. (1999). The relation between single crystal elasticity and the effective elastic behaviour of polycrystalline materials: Theory, measurement and computation. Modelling and Simulation in Materials Science and Engineering, 7(6), 909. 16. Arakere, N.  K., & Swanson, G. (2000). Effect of crystal orientation on fatigue failure of single crystal nickel base turbine blade superalloys. In ASME turbo expo 2000: Power for land, sea, and air. American Society of Mechanical Engineers. 17. Tang, W., Sanville, E., & Henkelman, G. (2009). A grid-based Bader analysis algorithm without lattice bias. Journal of Physics: Condensed Matter, 21(8), 084204. 18. Henkelman, G., Arnaldsson, A., & Jónsson, H. (2006). A fast and robust algorithm for Bader decomposition of charge density. Computational Materials Science, 36(3), 354–360. 19. Plimpton, S. (1995). Fast parallel algorithms for short-range molecular dynamics. Journal of Computational Physics, 117(1), 1–19. 20. Crocombette, J.-P., & Chartier, A. (2007). Molecular dynamics studies of radiation induced phase transitions in La 2 Zr 2 O 7 pyrochlore. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, 255(1), 158–165. 21. Hestenes, M.  R., & Stiefel, E. (1952). Methods of conjugate gradients for solving linear systems (Vol. 49). NBS.

46

3  Modeling of Mechanical Properties of Lanthanum Zirconate Crystal

22. Barsoum, M. W. (2002). Fundamentals of ceramics. CRC Press. 23. Shimamura, K., et  al. (2007). Thermophysical properties of rare-earth-stabilized zirconia and zirconate pyrochlores as surrogates for actinide-doped zirconia. International Journal of Thermophysics, 28(3), 1074–1084. 24. Liu, B., et al. (2007). Theoretical elastic stiffness, structure stability and thermal conductivity of La2Zr2O7 pyrochlore. Acta Materialia, 55(9), 2949–2957. 25. Fischer-Cripps, A. C., & Lawn, B. R. (1996). Stress analysis of contact deformation in quasi-­ plastic ceramics. Journal of the American Ceramic Society, 79(10), 2609–2618. 26. Zhang, J., et  al. (2014). Microstructural non-uniformity and mechanical property of air plasma-sprayed dense lanthanum zirconate thermal barrier coating. Materials Today: Proceedings, 1(1), 11–16. 27. Lilleodden, E. T., et al. (2003). Atomistic simulations of elastic deformation and dislocation nucleation during nanoindentation. Journal of the Mechanics and Physics of Solids, 51(5), 901–920. 28. Dupont, V., & Germann, T. C. (2012). Strain rate and orientation dependencies of the strength of single crystalline copper under compression. Physical Review B, 86(13), 134111.

Chapter 4

Modeling of Thermal Properties of Lanthanum Zirconate Crystal

4.1  Introduction of Thermal Properties Calculations TBCs are operated in high-temperature environment, so thermal properties are very important for TBC material. The primary requirement of TBCs for the gas turbine applications is TBCs’ low thermal conductivity [1]. Although there are several experimental methods to measure thermal conductivity, such as flash method and pulsed thermal imaging-multilayer analysis (PTI-MLA) method, there is no theoretical modeling calculation on LZ thermal conductivity. Actually, the intrinsic thermal conductivity of single crystal LZ is almost impossible to be accurately measured using experimental method due to the porous and fragile characteristic of the LZ powder, which includes a huge number of defects. In addition, experiment has certain limitations such as limited temperature range and sample size. However, theoretical calculation does not have temperature limitation. In this chapter, the specific heat of single LZ crystal was calculated based on the optimized LZ unit cell using DFT calculations. Then, we propose a novel image-­ based multi-scale simulation framework combining MD and FE methods to study the thermal conductivity of LZ thermal barrier coating. A reverse non-equilibrium molecular dynamics approach is first used to compute the temperature-dependent thermal conductivity of LZ single crystal. Then, the single crystal data is plugged into an FE model of a thermal barrier coating that is generated using SEM microstructures images. The predicted thermal conductivities from the FE model are compared against thermal conductivity experimental results using both flash laser and PTI-MLA techniques.

© Springer Nature Switzerland AG 2021 X. Guo et al., Novel Lanthanum Zirconate-based Thermal Barrier Coatings for Energy Applications, https://doi.org/10.1007/978-3-030-58695-9_4

47

48

4  Modeling of Thermal Properties of Lanthanum Zirconate Crystal

4.2  Thermodynamic Energy and Specific Heat of LZ The thermodynamic energies (Gibbs free energy, enthalpy, and entropy) and specific heat calculation of LZ unit cell were implemented by CASTEP code [2]. The thermodynamic energy and specific heat can be well described using a quasi-­ harmonic Debye model, in which the non-equilibrium Gibbs function G*(V, P, T) can be expressed as follows [3].

G V  x  , P, T   E  x   PV  x   Avib V  x  , T 



(4.1)

where x represents all the geometric information such as lattice constant, E(x) is the total energy of LZ unit cell, PV(x) represents the energy in constant hydrostatic pressure condition, and Avib is the vibrational Helmholtz free energy. E(V) and PV can be obtained by calculating DFT, so the next step is to calculate the vibrational contribution Avib using Debye model of the phonon density of states:



9     Avib  ,T   nkT   3 ln 1  e   / T  D     T  8 T





(4.2)

where Θ(V) is Debye temperature, n represents the number of atoms per formula unit, and D(y) is the Debye integral that is defined as: D  y 

y

3 x3 dx 3  x y 0 e 1

(4.3)

In Debye’s theory, the vibrations of solid are considered as elastic waves, so the Debye temperature of solids is related to average sound velocity. In addition, Debye temperature in a solid material is determined by phonon perturbation and lattice vibration, which can be computed as follows:





h 6 2V 1/ 2 n k





1/ 3

f  

B M

(4.4)

where M represents the molecular mass, B is the adiabatic bulk modulus, ν is Poisson ratio, and f(ν) is given by the following equation: 1/ 3



2/3 2 / 3 1

  2 1   1 1   f    3 2     3 1          3 1  2  

(4.5)

The phonon calculations in CASTEP were used to evaluate the enthalpy, entropy, free energy, and specific heat capacity of LZ unit cell as a function of temperature in the quasi-harmonic Debye approximation. The specific heat (Cv) can be expressed using the following equation [3]:

4.2  Thermodynamic Energy and Specific Heat of LZ



 3 / T   Cv  3nk  4 D   / T    / T  e 1  

49

(4.6)

The calculated thermodynamic energy curves are shown in Fig. 4.1, at temperatures of 0–1800 K. The entropy increases as temperature increases, which is consistent with the second law of thermodynamics. Then, enthalpy increases slowly and the free energy decreases as the temperature increases. The enthalpy indicates the amount of heat transfer into the crystal cell. Since the Gibbs free energy equals to the enthalpy minus the product of temperature and entropy, it is reasonable that the free energy decreases as temperature increases. The calculated specific heat (Cv) curve under constant pressure is shown in Fig. 4.2, in the temperature range of absolute zero to 1327 °C, and such wide temperature range is very difficult to measure by experiment. The value of specific heat increases as the temperature increases, especially in low temperature range it increases very rapidly. When the temperature reaches above 400  °C, the specific heat increases very slowly. The specific heat values are 0.4–0.467 J/(g °C) at the temperature range from 400 °C to 1400 °C. By comparison with the experimental data from Vassen’s work [4, 5] and calculated results from Chartier’s work [6], it can be concluded that the specific heat calculated by this work is accurate and reliable.

Fig. 4.1  Thermodynamic energy curves of LZ as function of temperature

50

4  Modeling of Thermal Properties of Lanthanum Zirconate Crystal

0.6

Specific heat (J/g°C)

0.5 0.4 0.3 Vassen's experiment

0.2

Chartier' s MD calculation

0.1 0 -273

This work

-73

127

327 527 727 Temperature (oC)

927

1127

1327

Fig. 4.2  Specific heat results in comparison to the literature

4.3  Thermal Conductivity of LZ 4.3.1  T  hermal Conductivity of Single Crystal LZ Using MD Simulation The MD method can be used to investigate the thermal conductivity. The thermal conductivity of single crystals can be calculated by two common molecular dynamics methods: direct method [7, 8] and Green-Kubo method [9, 10]. The direct method is a non-equilibrium molecular dynamics (NEMD) method which imposes a temperature gradient to the system. The Green-Kubo method is an equilibrium MD (EMD) method which uses current fluctuation to calculate the thermal conductivity according to the fluctuation-dissipation theorem [11]. Based on the NEMD, a more reliable method to compute thermal conductivity was developed, which is the reverse NEMD (RNEMD) method [12]. In RNEMD method, the Muller-Plathe algorithm [13] is applied to exchange kinetic energy between two atoms in different regions of the simulation box at every finite step to induce a temperature gradient in the system. It works by exchanging velocities between two atoms in different parts of the simulation cell. At set intervals, the velocity of the fastest atom in one region is replaced by the velocity of the slowest atom in another region and vice versa. Consequently, the first region becomes colder, whereas the second region increases in temperature. The system will be reacted by flowing energy from the hot to cold regions. Eventually, a steady state is established when the exchanged energy equilibrates the energy flowing back in a temperature gradient over the space between the two regions. This enables the thermal conductivity of a material to be calculated. The usual NEMD approach is to impose a temperature gradient on the system and measure the response as the resulting heat flux. In RNEMD using the Muller-Plathe

4.3  Thermal Conductivity of LZ

51

algorithm, the heat flux is imposed, and the temperature gradient is the system’s response. The advantage of RNEMD over traditional NEMD is that there are no artificial “temperature walls” in the simulated system, since these cause a fluid structure different from the bulk. Additionally, energy and momentum are conserved and there are no thermostat issues [12]. The reverse non-equilibrium molecular dynamics (RNEMD) method is used to predict temperature-dependent thermal conductivities of the single crystal LZ model in this work. The thermal conductivity model is assembled using the geometry-­ optimized LZ unit cell, which contained 2 × 2 × 30 unit LZ cell and had the dimension of 21.6 × 21.6 × 324 Å3 in x,y,z direction [14]. The thermal conductivity model (LZ supercell model) contains a total of 10,560 atoms, including 6720 O atoms, 1920 La atoms, and 1920 Zr atoms. The supercell model is sliced into 30 layers with equal thickness. A temperature decay constant of 0.1 per picoseconds is imposed in each layer. The thermal conductivity is calculated by dividing the energy flux by the temperature gradient [12]:



k

q T / z

(4.7)

where k is the thermal conductivity, q is the heat flux defined by the amount of heat (Q) transferred per unit time per unit area in heat transfer direction, and ∂T/∂z is the temperature gradient in the heat transfer direction. The universal force field was used in this work, which has a full coverage of the periodic table [15]. One of the calculated temperature distribution contours in the LZ supercell model is shown in Fig. 4.3. There are two high temperature hot zones at the ends due to the periodic boundary condition and a low temperature cold zone in the middle for generating a temperature gradient. The target temperature in Fig. 4.3 is 1273 K, which is the average temperature in the supercell [16]. The temperature gradient along the heat flux direction in the supercell is plotted in Fig. 4.4. Two linear temperature gradients were identified and the thermal conductivity was calculated using Eq. 4.7. Fig. 4.3 Temperature distribution in LZ single crystal supercell at 1273 K

52

4  Modeling of Thermal Properties of Lanthanum Zirconate Crystal

Fig. 4.4  Temperature distribution along the heat flux direction

Fig. 4.5  Temperature-dependent thermal conductivity of LZ single crystal calculated using RNEMD method

The calculated temperature-dependent thermal conductivity of LZ single crystal is shown in Fig. 4.5. The thermal conductivity values range from 1.25 W/m/K to

4.3  Thermal Conductivity of LZ

53

1.39 W/m/K at temperatures of 473–1273 K. It should be noted that thermal radiation effect at high temperatures is not considered in the RNEMD model. The LZ single crystal thermal conductivity predicted by us is lower than that by Schelling’s NEMD result (1.98 W/m/K at 1200 °C [7]), probably due to two factors. One is the method: Schelling used NEMD and we used RNEMD. The second is the different force fields used in the model. Schelling used Buckingham potentials and we used universal force field. Schelling’s single crystal result compared reasonably well with fully dense polycrystalline experimental data. However, the influence of microstructure in the polycrystalline was not considered in Schelling’s work. In our model, we also calculate temperature-dependent thermal conductivity of polycrystalline LZ using the FE model as discussed below.

4.3.2  T  hermal Conductivity of Polycrystalline LZ Coating By Using FE Method Finite element (FE) method can be used to simulate the heat conduction process of coating structures with cracks and pores [17]. Pore and crack morphology of thermal barrier coating is an important parameter affecting the mechanical and thermal properties [18, 19]. The quantitative imaging analysis method can be used to investigate the non-uniformity properties of the porous coating with polycrystalline microstructure [20, 21]. Based on the quantitative imaging analysis, the image-­ based FE method use scanning electron microscope (SEM) images to generate microstructures and import into an FE model [22]. Therefore, the image-based FE method is able to calculate the thermal conductivity of non-uniform porous polycrystalline coatings. Three representative SEM images of cross-sectional view of the porous LZ TBC sample with porosity of 13.61% are converted into binary images by using imaging processing software package ImageJ [23]. In the binary images, white color regions represent solid LZ coating phase, and black color regions are pores and cracks. The binary images then are processed with an FE software package, COMSOL Multiphysics [24], to automatically generate FE meshes. In the FE models, a constant temperature difference boundary condition is applied at the top (Ttop) and bottom (Tbot) surface of the system, with the average temperature (Ttop/2 + Tbot/2) as the target temperatures, 473–1273 K. The pores and cracks are filled with non-flowing air. Zero thermal conductivity is assumed for the pores and cracks because the thermal conductivity of non-flowing air is much smaller than that of the coating [25, 26]. For the LZ solid phase, the calculated temperature-­dependent thermal conductivities from LZ single crystal using the RNEMD method are used. For porous LZ polycrystalline coating samples, SEM images of cross-sectional views of three free-standing coating samples were used. A representative SEM image of LZ coating is shown in Fig. 4.6a. The simulated temperature contour in the sample is shown in Fig. 4.6b. As shown in Fig. 4.7, comparing with the flash method

54

4  Modeling of Thermal Properties of Lanthanum Zirconate Crystal

Fig. 4.6 (a) SEM image and (b) calculated temperature contours of LZ

Fig. 4.7  Thermal conductivity of LZ coating – FE method prediction and flash method

experiment results, the thermal conductivity values calculated by FE method are in good agreement with the experimental results. The cracks and pores in the coating clearly reduce the thermal conductivity of the coatings. The average thermal conductivity measured from the three LZ porous coating samples by using flash method is given in Fig.  4.7, which ranges from 0.44 to 0.62 W/m/K at temperatures of 300–1200 K. The measured thermal conductivity is relatively constant below 900 K and slightly increases above 900 K. As shown in Fig. 4.7, the thermal conductivity measured by using FE model is in good agreement with experimental measurement.

4.3  Thermal Conductivity of LZ

55

Fig. 4.8  LZ coating: (a) optical image, (b) thermal conductivity map, (c) product of heat capacity (ρc) map

Pulsed thermal imaging-multilayer analysis (PTI-MLA) method is recently developed as a new method to measure the thermal conductivity [27] to overcome the limitations in the flash method. For example, in the flash method, both sides of the specimens are needed to access. The sample surface is destructed by the laser flash. In addition, the flash method obtains a single averaged thermal conductivity of the sample, and it does not provide spatial distribution of the thermal conductivity [27]. In comparison, the PTI-MLA technique is a one-sided experimental system, which consists of a photographic flash lamp (source 6400; BALCAR, France), an infrared filter, and a mid-wavelength (3–5 μm) infrared camera (Phoenix, FLIR). PTI-MLA method is based on monitoring the surface temperature decay, after it is applied with a pulsed thermal energy that is transferred inside the sample gradually [27]. Figure 4.8 shows the optical image of the LZ coating specimen, its thermal conductivity map, and its heat capacity map, obtained by the pulsed thermal image-­ multilayer analysis method. The heat capacity measured is ~2.16  J/cm3/K.  The average thermal conductivity measured by the PTI-MLA technique is 0.55 W/m/K, which is also consistent with the results measured by flash method and calculated by FE model.

4.4  Summary In this chapter, some thermodynamic energies (Gibbs free energy, enthalpy, and entropy) and the specific heat of single crystal LZ were calculated using DFT calculations based on a quasi-harmonic Debye model. In addition, a novel image-based multi-scale simulation framework combining molecular dynamics and finite element calculations has been proposed to study the thermal conductivity of porous LZ coatings. Experimental validations include the flash method and pulsed thermal

56

4  Modeling of Thermal Properties of Lanthanum Zirconate Crystal

image-multilayer analysis technique to measure the thermal conductivity of coating. The main conclusions of this are summarized as follows: 1. The calculated specific heat capacity of single crystal LZ ranges from 0.4  to 0.467  J/g/K at a temperature range from 400 °C to 1400 °C.  The calculated results have a good accuracy when compared with the literature data. 2. The thermal conductivity of LZ single crystal calculated by RNEMD method ranges from 1.25 W/m/K to 1.39 W/m/K at a temperature of 473–1273 K. 3. The thermal conductivities of porous LZ polycrystalline calculated by image-­ based FE method and single crystal thermal conductivity data range from 0.46 to 0.59 W/m/K at temperatures of 473–1273 K. The FE data measured are in good agreement with the flash method, 0.44 to 0.62 W/m/K, at a temperature range of 300–1200 K, and the PTI-MLA technique, 0.55 W/m/K. The image-based multi-­ scale simulation framework proposed in this chapter provides a tool to design advanced coating systems.

References 1. Clarke, D., & Levi, C. (2003). Materials design for the next generation thermal barrier coatings. Annual Review of Materials Research, 33(1), 383–417. 2. Segall, M. D., et al. (2002). First-principles simulation: Ideas, illustrations and the CASTEP code. Journal of Physics: Condensed Matter, 14(11), 2717. 3. Blanco, M. A., Francisco, E., & Luaña, V. (2004). GIBBS: Isothermal-isobaric thermodynamics of solids from energy curves using a quasi-harmonic Debye model. Computer Physics Communications, 158(1), 57–72. 4. Vassen, R., et al. (2000). Zirconates as new materials for thermal barrier coatings. Journal of the American Ceramic Society, 83(8), 2023–2028. 5. Lehmann, H., et al. (2003). Thermal conductivity and thermal expansion coefficients of the lanthanum rare-earth-element zirconate system. Journal of the American Ceramic Society, 86(8), 1338–1344. 6. Chartier, A., et al. (2003). Atomistic modeling of displacement cascades in La2Zr2O7 pyrochlore. Physical Review B, 67(17), 174102. 7. Schelling, P. K., Phillpot, S. R., & Grimes, R. W. (2004). Optimum pyrochlore compositions for low thermal conductivity. Philosophical Magazine Letters, 84(2), 127–137. 8. Maiti, A., Mahan, G. D., & Pantelides, S. T. (1997). Dynamical simulations of nonequilibrium processes — Heat flow and the Kapitza resistance across grain boundaries. Solid State Communications, 102(7), 517–521. 9. Kubo, R. (1991). In M. Toda & N. Hashitsume (Eds.), Statistical physics II nonequilibrium statistical mechanics (2nd ed.). Berlin, Heidelberg: Springer Berlin Heidelberg. 10. Che, J., et al. (2000). Thermal conductivity of diamond and related materials from molecular dynamics simulations. The Journal of Chemical Physics, 113(16), 6888–6900. 11. Schelling, P. K., Phillpot, S. R., & Keblinski, P. (2002). Comparison of atomic-level simulation methods for computing thermal conductivity. Physical Review B, 65(14), 144306. 12. Müller-Plathe, F. (1997). A simple nonequilibrium molecular dynamics method for calculating the thermal conductivity. The Journal of Chemical Physics, 106(14), 6082–6085. 13. Müller-Plathe, F. (1999). Reversing the perturbation in nonequilibrium molecular dynamics: An easy way to calculate the shear viscosity of fluids. Physical Review E, 59(5), 4894–4898.

References

57

14. Guo, X., & Zhang, J. (2014). First principles study of thermodynamic properties of lanthanum zirconate. Materials Today: Proceedings, 1(1), 25–34. 15. Rappé, A. K., et al. (1992). UFF, a full periodic table force field for molecular mechanics and molecular dynamics simulations. Journal of the American Chemical Society, 114(25), 10024–10035. 16. Guo, X., et  al. (2016). Image-based multi-scale simulation and experimental validation of thermal conductivity of lanthanum zirconate. International Journal of Heat and Mass Transfer, 100, 34–38. 17. Wang, L., et al. (2016). Modeling of thermal properties and failure of thermal barrier coatings with the use of finite element methods: A review. Journal of the European Ceramic Society, 36(6), 1313–1331. 18. Wang, L., et  al. (2011). Influence of pores on the surface microcompression mechanical response of thermal barrier coatings fabricated by atmospheric plasma spray—Finite element simulation. Applied Surface Science, 257(6), 2238–2249. 19. Arai, M., Ochiai, H., & Suidzu, T. (2016). A novel low-thermal-conductivity plasma-sprayed thermal barrier coating controlled by large pores. Surface and Coatings Technology, 285, 120–127. 20. Zhang, J., et  al. (2014). Quantitative analysis of pore morphology in lanthanum zirconate thermal barrier coating. In Materials Science and Technology (MS&T) 2014. The Minerals, Metals & Materials Society (TMS), Pittsburgh, PA, USA 21. Zhang, J., et  al. (2014). Microstructural non-uniformity and mechanical property of air plasma-sprayed dense lanthanum zirconate thermal barrier coating. Materials Today: Proceedings, 1(1), 11–16. 22. Keyak, J. H., et al. (1990). Automated three-dimensional finite element modelling of bone: A new method. Journal of Biomedical Engineering, 12(5), 389–397. 23. ImageJ. Available from: https://imagej.nih.gov/ij/ 24. COMSOL. (2016). Available from: https://www.comsol.com/ 25. Wang, Z., et  al. (2003). Effects of pores and interfaces on effective properties of plasma sprayed zirconia coatings. Acta Materialia, 51(18), 5319–5334. 26. Nair, B., Singh, J., & Grimsditch, M. (2009). A model for residual stress evolution in air-­ plasma-­sprayed zirconia thermal barrier coatings. In 24th Annual Conference on Composites, Advanced Ceramics, Materials, and Structures-A: Ceramic Engineering and Science Proceedings. Wiley. 27. Sun, J. (2014). Pulsed thermal imaging measurement of thermal properties for thermal barrier coatings based on a multilayer heat transfer model. Journal of Heat Transfer, 136(8), 081601.

Chapter 5

Modeling of Mechanical Properties of Zirconia Top Coat and Bond Coat Interface

5.1  Introduction of Interface Modeling It is widely accepted that the mechanical properties at the ceramic-metal interface between the top and bond coats and/or between the top coat and the TGO layer have a primary influence on the lifetime of the TBCs in the thermomechanical environments. It is still challenging to directly examine the interfacial mechanical properties from experiments. Modeling and simulation are powerful tools as an alternative way to investigate the interfacial properties and decipher failure mechanisms [1]. The failure and spallation mechanisms of the TBCs have been discussed by Evans in the context of continuum mechanics [2, 3, 4]. He suggested that the delamination, typically observed in TGO layer or near the interface, is related to a significant residual stress gradient which amplifies the imperfections in TBCs. Cracks propagate when the residual tensile and/or shear stresses exceed the delamination toughness of the top bond coat interface. At the atomic level, the mechanical characteristics at the ceramic-metal interface are related to the intrinsic atomic properties. The idealized mechanical properties can be calculated, which offers insights into the complex interface systems. Guo et al. investigated the mechanical properties of Ni (111)/α-Al2O3 (0001) interface, and calculated the theoretical shear strength and unstable stacking energy, using the first principles calculations [5]. Since LZ, 8YSZ, and NiCrAlY bond coat structure is too complicated to calculate using density functional theory (DFT) method, the ZrO2/Ni interface structure was implemented as a simplified top and bond coat model. This simplification is reasonable because the main composition of 8YSZ and LZ is ZrO2, and Ni is the main composition of bond coat. Christensen et al. studied the adhesion energy of ZrO2(111)/Ni(111) interface using the ultrasoft pseudopotential within the density functional theory [6]. In spite of previous efforts, the exact mechanical properties of ZrO2/Ni interface are still not well examined with theoretical calculation approaches. Specifically, the interfacial mechanical behaviors under tensile and shear stresses are not available. The difficulty of such studies primarily stems from the complexity of the interface © Springer Nature Switzerland AG 2021 X. Guo et al., Novel Lanthanum Zirconate-based Thermal Barrier Coatings for Energy Applications, https://doi.org/10.1007/978-3-030-58695-9_5

59

60

5  Modeling of Mechanical Properties of Zirconia Top Coat and Bond Coat Interface

structure, which requires minimizing the misfit between different crystal surfaces and intensive calculations involved in the interfacial tensile and shear deformation simulations. In this chapter, we discuss the DFT and molecular dynamics (MD) tensile and shear calculations of ZrO2 (111)/Ni (111) interface. In the DFT modeling, an interface model described in Ref. [6] is used to compare and/or partially validate the model by calculating the adiabatic work of adhesion. For mechanical property calculations, two models with Ni thickness of 1 and 3 atomic layers, respectively, are used to investigate the effect of interface thickness on the mechanical properties. The crystallographic orientation (111) is considered since the primary slip system in face-centered cubic (fcc) metal crystals, such as Ni, is {111}, with {111} being the secondary slip system [7]. Although the actual slip systems of the ZrO2/Ni interface may be more complicated, both {111} and {111} should be the major slip systems. Therefore, the tensile stress-strain curve in direction is calculated, and the shear stress-strain curves along {111} and {111} directions are computed. The toughness and elastic modulus, Young’s modulus or shear modulus, are also calculated. Finally, the Bader charge analyses are conducted to explain the observed interfacial mechanical properties. In the MD model, electronic charge was considered in the atomic force field. The tensile simulation in [111] direction and shear simulations in {111} and {111} direction were conducted using MD method.

5.2  DFT Methods of ZrO2/Ni Interface Model 5.2.1  ZrO2 (111)/Ni (111) Interface Model The constructed ZrO2 (111)/Ni (111) interface atomistic models are shown in Fig. 5.1. The cubic zirconia (c-ZrO2) has the fluorite crystal structure (space group Fm3m) and Ni has a face-centered cubic (fcc) crystal structure (space group Fm3m). Both c-ZrO2 and Ni small portions are cut from their bulk counterpart on (111) plane. To match the interface with minimal lattice misfit, the c-ZrO2 and Ni small portions are rotated according to Ref. [6]. As shown in Fig.  5.1, the interface model contains two layers of ZrO2 (111) atoms, and 1 or 3 layers of Ni (111) atoms [8]. Because the O atoms and Zr atoms are not positioned in the same horizontal plane, the termination of the ZrO2 (111) surface may result in dipole moment perpendicular to the interface. Therefore, symmetric models are built to screen out the dipole interactions. Two ZrO2 (111)/Ni (111) interfaces are in this interface model, corresponding to a sandwich structure. This model with the interface is extended periodically in three dimensions, due to the periodical boundary condition. To calculate adiabatic work of adhesion, ZrO2 (111)/Ni (111) interface slabs are built with a vacuum layer thickness of 10  Å, which is large enough to eliminate the interaction between each interface film layer.

5.2  DFT Methods of ZrO2/Ni Interface Model

61

Fig. 5.1 ZrO2 (111)/Ni (111) interface models with different Ni atomic layers: (a) side view and (b) top view with 3 layers of Ni, and (c) side view and (d) top view of 1 layer of Ni. Green, red, and black balls are Zr, O, and Ni atoms, respectively (same coloring schemes are used in all of the DFT modeling figures afterward)

The first principles calculations in this chapter are carried out using VASP [9– 11], based on the density functional theory (DFT) [12, 13]. The projector augmented wave (PAW) method of generalized gradient approximation (GGA) functional is adopted to specify the exchange-correlation potential. A periodic supercell regime is used and k-point of the Brillouin zone is conducted using 3 × 3 × 1 Monkhorst-­ Pack scheme. A conjugate-gradient algorithm is used to relax the ions into its instantaneous ground state. The plan-wave cutoff energy is 400  eV.  The energy relaxation criterion for the electron is 10−5 eV for self-consistency. The total forces are smaller than 0.01 eV/A in the ionic relaxation convergence criterion.

62

5  Modeling of Mechanical Properties of Zirconia Top Coat and Bond Coat Interface

5.2.2  Adiabatic Work of Adhesion In order to compare or partially validate the interface model, due to limited literature data, adiabatic work of adhesion (Wadh) is calculated using the atomistic model described in Ref. [6]. Wadh is the most commonly used property to describe the adhesion characteristics, which can be described using the following equation [6]:



Wadh 

N E0, Ni  E0N, Zro2  EZro 2 , Ni

A



(5.1)

where E0, Ni and E0,ZrO2 are the total energies of the relaxed Ni and ZrO2 surfaces, respectively. EZrO2 , Ni is the total energy of the relaxed ZrO2/Ni interface structure. A is the area of the interface and N is the number of the ceramic layers [6].

5.2.3  S  tress-Strain Behaviors in Tensile and Shear Deformations For tensile deformation, the calculations are conducted by extending the lattice parameters of the interface model in [111] direction, which is perpendicular to the interface. All ions in the interface model are relaxed, and the volume and the shape of the interface unit cell are also optimized during the stress tensor calculation process. Similarly, for shear deformations, the shear stress is calculated by accumulating the shear strain in {111} or {111} direction [14].

5.2.4  Bader Charge Analysis To explain the calculated stress-strain behaviors, the Bader method is used to calculate the charge transfer numbers and election density distributions [15, 16, 17]. The charge transfer results are processed by calculating the average charge difference between O and Ni ions. This is because, in the ZrO2/Ni interface models, the bonds are formed through the Zr and Ni atoms losing their electrons, and the O atoms gaining their electrons.

5.3  DFT Interface Modeling Results and Discussion 5.3.1  Adiabatic Work of Adhesion The relaxed ZrO2 /Ni interface models for adiabatic work of adhesion calculations are shown in Fig. 5.2. In this work, all the atoms except the two bottom Ni layers (which are away from the interface) are relaxed to allow to reach their equilibrium.

5.3  DFT Interface Modeling Results and Discussion

63

Fig. 5.2  Relaxed ZrO2 (111)/Ni (111) interface models with a vacuum layer above the slabs: (a) 1-layer ZrO2, (b) 2-layer ZrO2, and (c) 3-layer ZrO2

Comparing with the structures given in Ref. [6], the relaxed atomic structures in this study are slightly different. This is due to limited information regarding computation details in Ref. [6]. In this chapter, the calculated Wadh value of the interface with 1-layer ZrO2 is 629  mJ/m2, which is greater than those of 2- and 3-layer ZrO2 (554  mJ/m2 and 296 mJ/m2, respectively). In Ref. [6], the Wadh values are 2011, 1308, and 995 mJ/m2 for 1-, 2-, and 3-layers ZrO2, respectively. Although our calculated values are lower than those reported in Ref. [6] by ~60%, our calculated values follow the same trend as Ref. [6], i.e., a thicker ZrO2 layer corresponds to a lower adhesion energy.

5.3.2  S  tress-Strain Behaviors in Tensile and Shear Deformations The atomic configurations of the relaxed tensile models along [111] direction are shown in Figs. 5.3 and 5.4, for the Ni slabs of 1 and 3 layers, respectively. The tensile stress-strain curves of ZrO2 (111)/Ni (111) interface with Ni slabs of 1 and 3 layers are shown in Fig. 5.5. As shown in Fig. 5.5, a thinner (1 layer) Ni layer has almost double Young’s modulus (139.9 GPa) and higher ultimate tensile

64

5  Modeling of Mechanical Properties of Zirconia Top Coat and Bond Coat Interface

Fig. 5.3  Tensile deformation models with 1 layer of Ni: (a) initial configuration, (b) strain of 0.051, (c) strain of 0.105, and (d) atom displacement vector between strain 0.105 and initial steps

Fig. 5.4  Tensile deformation models with 3 layers of Ni: (a) initial configuration, (b) strain of 0.072, (c) strain of 0.138, and (d) atom displacement vector between strain 0.138 and initial steps

strength (11.6 GPa) than the 3-layer Ni layer (60.2 GPa and 7.9 GPa, respectively) (see Table 5.1 for a complete summary of calculated mechanical properties, including elastic modulus, ultimate tensile/shear strength, and toughness). The layer-­ thickness dependence is consistent with work of adhesion values given in Sect. 5.3.1, i.e., a thin Ni layer interface model has higher strength or work of adhesion. In terms of deformation strain, the 3-layer Ni interface is more ductile as illustrated with larger tensile strain. This can be interpreted by the atom displacement vectors between the final and initial steps of the nanoscale tensile calculation, as shown in Figs. 5.3d and 5.4d. As shown in Fig. 5.4d, the atom displacement of Ni atoms at the interface is larger than that of Zr and O atoms in 3-layer Ni interface model, indicating that most of the deformation occurs among Ni layers in the 3-Ni-layer model. In

5.3  DFT Interface Modeling Results and Discussion

65

Fig. 5.5  Tensile stress-strain curves of ZrO2(111)/Ni (111) interfaces with Ni slabs of 1 and 3 layers Table 5.1  Calculated elastic modulus, ultimate strength, and toughness Deformation mode, stress direction, and number of Ni layers Tensile [111] 1-layer Ni Tensile [111] 3-layer Ni Shear {111} 1-layer Ni Shear {111} 3-layer Ni

Elastic Ultimate tensile/shear modulus (GPa) strength (GPa) 139.9 11.6 60.2 7.9 43.9 7.9

Toughness (MJ/m3) 0.728 0.486 1.040

30.4

3.0

1.038

Shear {111} 1-layer Ni

30.9

6.0

0.166

Shear {111} 3-layer Ni

17.3

1.8

0.096

the 1-layer Ni interface model, the displacement of Ni, Zr, and O atoms is arbitrarily distributed, but the total outcome of these displacements extends the model in tensile direction. The 3-layer Ni interface possesses more deformation compatibility than 1-layer Ni interface, suggesting that a thick Ni interface can provide extra deformation to accommodate tensile strain. Shear Deformations Along {111} and {111} Directions The atomic configurations of 1-layer Ni slab model during shear deformations along {111} and {111} directions are shown in Figs.  5.6 and 5.7, respectively. The calculated shear stress-strain curves are given in Fig. 5.8. Both the shear modulus (43.9 GPa) and ultimate shear strength (7.9 GPa) along {111} direction are greater than those along {111} direction (30.9 GPa for shear

66

5  Modeling of Mechanical Properties of Zirconia Top Coat and Bond Coat Interface

Fig. 5.6  Shear deformation model with 1-layer Ni along {111} direction after relaxation: (a) initial position, (b) strain 0.126, (c) strain 0.230, and (d) atom displacement vectors between strain 0.230 and initial steps

Fig. 5.7  Shear deformation model with 1-layer Ni along {111}direction after relaxation: (a) initial position, (b) strain 0.126, (c) strain 0.267, and (d) atom displacement vectors between strain 0.267 and initial step

modulus and 6.0  GPa for ultimate shear strength). Therefore, {111} is a favorable shear slip system in this ZrO2 (111)/Ni (111) interface system, which is different from pure Ni. It is also noted that the ductility of {111} measured by strain, 0.23, is lower than that of {111}, 0.27. The shear deformations along {111} and {111} directions of the 3-layer Ni model are shown in Figs. 5.9 and 5.10, respectively. The calculated shear stress-strain curves are given in Fig. 5.11. Similar to the 1-layer Ni model, for the 3-layer Ni, both the shear modulus (30.4 GPa) and ultimate shear strength (3.0 GPa) along {111} direction are greater than those along {111} direction (17.3 GPa for shear modulus and 1.8 GPa for ultimate shear strength). Therefore, {111} is again a favorable shear slip system. Again, the ductility of {111} measured by strain, ~0.10, is lower than that of {111}, ~0.11.

5.3  DFT Interface Modeling Results and Discussion

67

Fig. 5.8  Calculated shear stress-strain curves of 1-layer Ni shear deformation model

Fig. 5.9  Shear deformation model with 3-layer Ni along {111} direction after relaxation: (a) initial position, (b) strain 0.051, (c) strain 0.105, and (d) atom displacement vectors between strain 0.105 and initial steps

The atom displacement vectors between the final and initial steps of shear deformation models are plotted in Figs.  5.6d, 5.7d, 5.9d, and 5.10d. In the 1-layer Ni interface models, there are no obvious differences for the displacement vectors between Ni, Zr, and O atoms, as shown in Figs. 5.6d and 5.7d. However, the displacement directions between upper and lower ZrO2/Ni interfaces are in opposite trend, which properly illustrates the shear deformation. As shown in Figs. 5.9d and 5.10d, the 3-layer Ni interface models show larger displacement in the Ni layers than in the ZrO2 layers, indicating that the Ni layer provides the most deformation in these interface models.

68

5  Modeling of Mechanical Properties of Zirconia Top Coat and Bond Coat Interface

Fig. 5.10  Shear deformation model with 3-layer Ni along {111}direction after relaxation: (a) initial position, (b) strain 0.062, (c) strain 0.116, and (d) atom displacement vectors between strain 0.116 and initial steps

Fig. 5.11  Calculated shear stress-strain curves of 3-layer Ni shear deformation model

Comparing with the 1-layer Ni shear deformation model, the shear moduli and strengths in the 3-layer model are much lower by 40%. This is because the shear deformation is primarily achieved by the deformation of Ni atoms. A thicker Ni layer allows to deformation at lower stress level. In addition to stress, toughness can be used to measure the interfacial strength in large deformations. Toughness is calculated by integrating the area below the stress-­ strain curve. As shown in Table  5.1, the 1-layer Ni interface model has higher

5.3  DFT Interface Modeling Results and Discussion

69

toughness than the 3-layer Ni model for both tensile and shear deformations. This is also consistent with the results from the adiabatic work of adhesion given in Sect. 5.3.1. In addition, the toughness in {111} system is higher than that of {111} direction for both 1-layer and 3-layer Ni models, primarily due to higher shear modulus and ultimate shear strength in the {111} system. It also suggests that the {111} direction is stronger than the {111} direction during interface fracture. Since most interfacial deformation is achieved by Ni atoms, it is worthy to compare the interfacial models with pure Ni. Ogata et al. calculated the shear stress-­ strain curves of pure Ni in {111} direction using the DFT calculations [18]. The calculated ultimate strength is 5.1 GPa. It is similar to the 1-layer Ni interface model in this work, 6.0  GPa. However, the shear modulus in Ogata’s work is ~60.3 GPa, which is larger than that in this work, 30.9 GPa. The Young’s modulus of polycrystalline Ni (190~220 GPa) from experiment [19] is also much higher than that of ZrO2/Ni interface calculated in this work (139.9 GPa for 1-layer and 60.2 GPa for 3-layer Ni). Both elastic modulus and ultimate strength values decrease as Ni layer thickness increases. These comparisons suggest that the strength of the ZrO2/ Ni interface is substantially different from its pure component and is determined by the Ni layer thickness.

5.3.3  Charge Density and Bader Charge Analyses The contours of charge density distribution are plotted in Fig. 5.12. As shown in the figure, it is clear that the 1-layer Ni interface models (Fig.  5.12a–c) have much strong O-Ni bonds than the 3-layer Ni models (Fig. 5.12d–f). This is the reason why the 1-layer Ni interface models have higher elastic modulus and strength. The electron localization function (ELF) can be used to describe the electron localization status and bonding behaviors, which depends on the electron density, gradient, and the kinetic energy density [20, 21]. Typically the EFL values vary from 0 to 1, corresponding to from vacuum to perfect localization of the electrons. A higher ELF value in the ELF graph indicates that the electrons are more localized. Metallic bonding in the ELF graph typically shows electron vacuum near the atom nucleus and a relative high electron localization value at area far from the atom nucleus. Figure 5.13 shows the ELF graph of 1-layer and 3-layer Ni shear {111} interface models; ELF graph in other tensile and shear cases shows similar trend. The ELF of 3-layer Ni model shows electron vacuum near the Ni atom; however, it has a comparatively higher electron localization value at space far from the Ni atom. The Ni layers in the interface model show a typical metallic bonding characteristics, which can accommodate the deformation during the tensile and shear process. This is consistent with the above discussions that thicker Ni layer results in lower elastic modulus and lower ultimate strength. As shown in Fig. 5.13, the O and Zr atoms have higher ELF values than the Ni atoms in both 1-layer and 3-layer Ni interface models. The chemical bonding

70

5  Modeling of Mechanical Properties of Zirconia Top Coat and Bond Coat Interface

Fig. 5.12  Charge density distributions in logarithmic scale: (a) tensile with 1-layer Ni, (b) shear {111} with 1-layer Ni, (c) shear {111} with 1-layer Ni, (d) tensile with 3-layer Ni, (e) shear {111} with 3-layer Ni, (f) shear {111} with 3-layer Ni

between the O and Ni atoms at the interface shows an ionic bonding feature. In addition, both ELF graphs show a delocalized electron gap between ZrO2 and Ni layers, suggesting that the ionic bonding might be weaker than the metallic bonding in Ni layer, which has higher ELF value than ionic bonding. During the tensile and shear processing, the ionic bonding becomes weaker and weaker, until it breaks. Because 1-layer Ni models have stronger ionic bonding than the 3-layer Ni models, as shown in Fig. 5.12 of the charge density distribution contour, the 1-layer Ni interface models possess higher ultimate strength than their 3-layer Ni counterpart. To get more quantitative measurements about bond characteristics, the average Bader charge numbers, including O, Zr, Ni ions, and the difference between O and

5.3  DFT Interface Modeling Results and Discussion

71

Fig. 5.13  ELF graph in the (a) shear {111} with 1-layer Ni interface model and (b) shear {111} with 3-layer Ni interface model Table 5.2  Average Bader charge number (e) Deformation mode, stress direction, and number of Ni layers Tensile 1-layer Ni Tensile 3-layer Ni Shear {111} 1-layer Ni Shear {111} 3-layer Ni Shear {111} 1-layer Ni Shear {111} 3-layer Ni

Zr −2.457 −2.459 −2.451

Ni −0.050 −0.019 −0.045

Difference between O and Ni 1.308 1.283 1.297

1.266 −2.462

−0.019

1.285

1.257 −2.460

−0.046

1.303

1.265 −2.460

−0.019

1.284

O 1.258 1.264 1.252

Ni ions for both tensile and shear deformations, are summarized in Table 5.2. It is clear that, for the 1-layer Ni interface models, both the O and Ni differences and the average Bader charge difference between O and Ni ions are consistently larger than those of 3-layer Ni interface, for both tensile and shear cases. The Bader charge analysis results are also consistent with the charge density distributions shown in

72

5  Modeling of Mechanical Properties of Zirconia Top Coat and Bond Coat Interface

Fig. 5.12. A larger average Bader charge difference indicates stronger interaction between O and Ni or more interaction between ZrO2 and Ni. This explains the higher ultimate strength and elastic modulus in the thin Ni-layer interface than in the thick Ni-layer interface, as listed in Table 5.1. From the calculated mechanical properties of ZrO2/Ni ceramic-metal interface, the layer thickness of bond coat film, NiCrAlY, at the interface makes a major impact on the coating’s mechanical behavior. Typically, fracture or delamination in as-sprayed TBC system occurs near the interfaces between the top and bond coats [22, 23]. Therefore, the mechanical properties near the interface are important to enhance the lifetime performance of TBC system. Higher toughness and elastic modulus at the interface enhance the ability of fracture resistance to impede crack propagations in the ceramic top coat near the interface. From the theoretical analyses of this work, the ultimate tensile and shear strength are decreased with increase of the bond coat film thickness at the interface, which means a thicker thickness of bond coat in TBC system corresponds to a weaker adhesion strength. On the other hand, a thin bond coat film will deteriorate the oxidation resistance of the TBC system. Therefore, the thickness of bond coat should be properly optimized to design and fabricate robust TBC systems.

5.4  M  D Tensile and Shear Simulations of ZrO2/Ni Interface Model The DFT calculation is performed on the ground energy state of the atomic system, which means that the DFT theoretical results are limited at very low temperature (close to 0  K). Therefore, MD method was performed to simulate the nanoscale tensile and shear process of ZrO2/Ni interface model at high temperature. LAMMPS code was used to conduct all of the MD calculations in this work. The charge-­ optimized many body (COMB3) potential was applied to describe the interatomic force field of ZrO2/Ni interface [23]. In COMB potential series (including both COMB and COMB3), the total potential energy Utot of system is described by the following equation:

U tot  r ,q   U es  r ,q   U short  r ,q   U vdW  r   U corr  r 



(5.2)

where r represents the coordinate array of the system and q represents the charge array. Ues represents the electrostatic energies that include the energies to form a charge on an atom, the charge-nuclear interaction, the charge-charge interaction, and the energies related with atomic polarizability. Ushort is the short-range interaction energies that describe the pairwise repulsive and attractive potentials. The long-­ range van der Waals interactions (UvdW) are described using Lennard-Jones expression. The energy corrections term (Ucorr) is used to optimize the total energy. To validate the COMB3 potential in ZrO2/Ni interface system, the lattice constants of ZrO2 unit cell and Ni unit cell were calculated using MD model,

5.4  MD Tensile and Shear Simulations of ZrO2/Ni Interface Model

73

respectively. Cubic ZrO2 with a space group Fm3m and cubic Ni of the same space group are used to calculate the lattice constant. The DFT calculation methods are similar as the one used in LZ unit cell model, except different cutoff energy and Brillouin zone k-point scheme. The cutoff energies for ZrO2 and Ni were 500 eV and 300 eV, respectively, and the 5 × 5 × 5 Monkhorst-Pack scheme was used for both ZrO2 and Ni as Brillouin zone k-point. Table 5.3 lists the DFT and MD calculation results of the ZrO2 and Ni lattice constants. The error differences between MD results and DFT results for ZrO2 and Ni are 1.478% and 0.378%, respectively, suggesting that the COMB3 potential is reliable on single material ZrO2 or Ni MD models. The ZrO2(111)/Ni(111) interface MD models were built based on the DFT model as described in Sect. 5.2. DFT Methods of ZrO2/Ni Interface Model, and the lattice mismatch was only about 2.68%. The work of adhesion of ZrO2(111)/Ni(111) interface was calculated according to Eq. (5.1) using MD method and it was compared with DFT results. The calculated work of adhesion results are listed in Table 5.3 and plotted in Fig. 5.14. The MD-calculated work of adhesion result has the same trend as DFT results and a moderate accuracy, which is acceptable in the interface simulations. Table 5.3  DFT- and MD-calculated lattice constant and work of adhesion results Lattice constant ZrO2 (Å) Lattice constant Ni (Å) Work of adhesion with 1-layer ZrO2 (mJ/m2) Work of adhesion with 2-layer ZrO2 (mJ/m2) Work of adhesion with 3-layer ZrO2 (mJ/m2)

DFT-calculated result 5.146 3.514 629 554 296

MD-calculated result 5.070 3.527 751 602 378

Fig. 5.14  Comparison of work of adhesion between DFT and MD models

74

5  Modeling of Mechanical Properties of Zirconia Top Coat and Bond Coat Interface

The tensile stress-strain curves of ZrO2/Ni interface model were calculated using the same scheme as LZ model. However, the shear stress and strain values here were calculated from the single direction pure shear deformation. Therefore, the shear strain was represented by the deformation angle, and shear stress was calculated from the elastic constants. The MD models of ZrO2(111)/Ni(111) interface before and after the tensile simulation are shown in Fig. 5.15. The periodic boundary condition was applied in all three dimensions of this tensile model, so the ZrO2-Ni-ZrO2 sandwich model was built to match the periodic condition in the tensile direction. As shown in Fig. 5.15b, the Ni layer in the model with 0.16 tensile strain did not keep the initial lattice array, suggesting that the primary deformation came from the Ni layer. The MD ZrO2(111)/Ni(111) interface model used in the shear simulation of {111} direction at 300 K is shown in Fig. 5.16. The non-periodic boundary conduction was applied in the shear direction and the periodic boundary condition was used in the other two dimensions. The tensile stress-strain curves of ZrO2(111)/Ni(111) interface model in [111] direction at 300 K are shown in Fig. 5.17. The tensile curve shows yield stage and reaches its ultimate strength of 9.62 GPa at a strain of 0.072. The ultimate strength is in the same range as DFT-calculated results. The undulation of the curve is due to the limitation of the atom numbers. The average Young’s modulus is 211.27 GPa, which is larger than the DFT results as listed in Table 5.1 (139.9GPa in 1-Ni-layer interface model and 60.3 GPa in 3 layer-Ni interface model). The shear stress-strain curves in {111} and {111} direction at 300 K are shown in Fig. 5.18. These two curves share the similar shearing stress-­ strain trend. The ultimate shear strength in {111} direction is ~9 GPa at a shear strain of ~0.35. The average shear modulus in {111} is ~29.39 GPa.

Fig. 5.15 ZrO2(111)/Ni(111) interface MD model in tensile simulations at 300 K (a) initial position with 0 strain, (b) the position of tensile strain 0.16. Green, red, and gray balls represent Zr, O, and Ni atoms, respectively (same coloring schemes are used in all of the MD figures afterward)

5.4  MD Tensile and Shear Simulations of ZrO2/Ni Interface Model

75

Fig. 5.16  MD ZrO2(111)/ Ni(111) interface model in shear simulation in {111} direction: (a) initial position with 0 strain and (b) the position of shear strain 0.45

Fig. 5.17  Tensile stress-strain curves of ZrO2/Ni interface at 300 K

These two curves share a similar shearing stress-strain trend. The ultimate shear strength in {111} direction is ~12 GPa at a shear strain of ~0.35. The average shear modulus in {111} is ~34.55 GPa. The ultimate strength is larger than DFT-calculated results, as listed in Table 5.1. However, the shear modulus is very close to the DFT-calculated results. The ultimate strength differences between DFT and MD results are due to the dimension scale difference in the model. In addition,

76

5  Modeling of Mechanical Properties of Zirconia Top Coat and Bond Coat Interface

the MD calculation is performed in a dynamic equilibration, and the DFT calculations are conducted in the static equilibration of each tensile steps.

5.5  Summary The ideal tensile strength and shear strength of ZrO2/Ni ceramic-metal interface are calculated using the DFT and MD methods. The major conclusions are summarized below. For tensile deformation in [111] direction, the Young’s moduli of the 1-layer Ni and 3-layer Ni metal-ceramic models are 139.9 GPa and 60.2 GPa, respectively; the ultimate tensile strengths are 11.6 GPa and 7.9 GPa, respectively; the toughnesses are 0.728 MJ/m3 and 0.486 MJ/m3, respectively. The 1-layer Ni model shows higher mechanical strength than the 3-layer Ni model in tensile deformation. For shear deformation in {111} system, the shear moduli of the 1-layer Ni and 3-layer Ni M-C models are 43.9 GPa and 30.4 GPa, respectively; the ultimate shear strengths are 7.0 GPa and 3.0 GPa, respectively; the toughnesses are 1.040 MJ/ m3 and 1.038 MJ/m3, respectively. The 1-layer Ni model shows higher mechanical strength than the 3-layer Ni model in shear deformation. For shear deformation in {111} system, the shear moduli of the 1-layer Ni and 3-layer Ni metal-ceramic models are 30.9 GPa and 17.3 GPa, respectively; the ultimate shear strengths are 6.0 GPa and 1.8 GPa, respectively; the toughnesses are 0.166 MJ/m3 and 0.096 MJ/m3, respectively.

Fig. 5.18  Shear stress-strain curves of ZrO2/Ni interface at 300 K

References

77

Both charge distribution and Bader charge analyses show that the 1-layer Ni ceramic-metal model has stronger interaction between Ni and O ions than the 3-layer Ni model, which explains the higher mechanical properties in 1-layer Ni model from the calculated tensile and shear stress-strain results. From the theoretical analyses of this work, the ultimate tensile and shear strength are decreased with the increase of the bond coat film thickness at the interface, which means a thicker thickness of bond coat in TBC system corresponds to a weaker adhesion strength. On the other hand, a thin bond coat film will deteriorate the oxidation resistance of the TBC system. Therefore, the thickness of bond coat should be properly optimized to design and fabricate robust TBC systems.

References 1. Finnis, M. (1996). The theory of metal-ceramic interfaces. Journal of Physics: Condensed Matter, 8(32), 5811. 2. Evans, A. G., et al. (2001). Mechanisms controlling the durability of thermal barrier coatings. Progress in Materials Science, 46(5), 505–553. 3. Mumm, D., & Evans, A. (2000). On the role of imperfections in the failure of a thermal barrier coating made by electron beam deposition. Acta Materialia, 48(8), 1815–1827. 4. Evans, A., & Hutchinson, J. (2007). The mechanics of coating delamination in thermal gradients. Surface and Coatings Technology, 201(18), 7905–7916. 5. Guo, X., & Shang, F. (2012). Shear strength and sliding behavior of Ni/Al2O3 interfaces: A first-principle study. Journal of Materials Research, 27(09), 1237–1244. 6. Christensen, A., & Carter, E. A. (2001). Adhesion of ultrathin ZrO2 (111) films on Ni (111) from first principles. Journal of Chemical Physics, 114(13), 5816–5831. 7. Arakere, N.  K., & Swanson, G. (2000). Effect of crystal orientation on fatigue failure of single crystal nickel base turbine blade superalloys. In: ASME turbo expo 2000: Power for land, sea, and air. American Society of Mechanical Engineers. 8. Guo, X., et  al. (2016). Ideal tensile strength and shear strength of ZrO2(111)/Ni(111) ceramic-metal Interface: A first principle study. Materials & Design, 112, 254–262. 9. Kresse, G., & Hafner, J. (1993). Ab initio molecular dynamics for liquid metals. Physical Review B, 47(1), 558. 10. Kresse, G., & Furthmüller, J. (1996). Efficiency of ab-initio total energy calculations for metals and semiconductors using a plane-wave basis set. Computational Materials Science, 6(1), 15–50. 11. Kresse, G., & Furthmüller, J. (1996). Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set. Physical Review B, 54(16), 11169–11186. 12. Hohenberg, P., & Kohn, W. (1964). Inhomogeneous electron gas. Physical Review, 136(3B), B864. 13. Kohn, W., & Sham, L. J. (1965). Self-consistent equations including exchange and correlation effects. Physical Review, 140(4A), A1133–A1138. 14. Zhang, M., et  al. (2014). Hardness of FeB4: Density functional theory investigation. The Journal of Chemical Physics, 140(17), 174505. 15. Tang, W., Sanville, E., & Henkelman, G. (2009). A grid-based Bader analysis algorithm without lattice bias. Journal of Physics: Condensed Matter, 21(8), 084204. 16. Henkelman, G., Arnaldsson, A., & Jónsson, H. (2006). A fast and robust algorithm for Bader decomposition of charge density. Computational Materials Science, 36(3), 354–360. 17. Bader, R. F. (1985). Atoms in molecules. Accounts of Chemical Research, 18(1), 9–15.

78

5  Modeling of Mechanical Properties of Zirconia Top Coat and Bond Coat Interface

18. Ogata, S., et  al. (2004). Ab initio study of ideal shear strength. In IUTAM symposium on mesoscopic dynamics of fracture process and materials strength. Springer. 19. Ledbetter, H., & Reed, R. P. (1973). Elastic properties of metals and alloys, I. Iron, nickel, and Iron-nickel alloys. Journal of Physical and Chemical Reference Data, 2(3), 531–618. 20. Becke, A.  D., & Edgecombe, K.  E. (1990). A simple measure of electron localization in atomic and molecular systems. The Journal of Chemical Physics, 92(9), 5397–5403. 21. Silvi, B., & Savin, A. (1994). Classification of chemical bonds based on topological analysis of electron localization functions. Nature, 371(6499), 683–686. 22. Guo, X., et al. (2016). Thermal properties, thermal shock, and thermal cycling behavior of lanthanum zirconate-based thermal barrier coatings. Metallurgical and Materials Transactions E, 3(2), 64–70. 23. Liang, T., et al. (2013). Classical atomistic simulations of surfaces and heterogeneous interfaces with the charge-optimized many body (COMB) potentials. Materials Science and Engineering: R: Reports, 74(9), 255–279.

Chapter 6

Concluding Remarks and Future Research Directions

6.1  Concluding Remarks Lanthanum zirconate is proposed as a promising thermal barrier coating material due to its outstanding phase stability at high temperature and lower thermal conductivity than the commercial 8YSZ TBCs material. In this book, layered LZ TBC (SCL LZ coatings with porosity range from 11.81% to 16.52% and DCL LZ/8YSZ coatings with LZ porosity of 11.54%) and composite LZ/8YSZ TBCs were designed and deposited using APS method. The physical, mechanical, and thermal properties of LZ coating were investigated through experiments and multi-scale modeling techniques. The thermal cycling tests of LZ-based coatings were performed to investigate the thermal durability performance. The tensile and shear nanoscale simulations of interface model were conducted to give an insight of the TBCs delamination mechanism. The primary conclusions of this book are summarized as follows. 1. The mechanical properties of layered LZ-based TBCs, such as hardness and Young’ s modulus, were investigated using Vickers microhardness and nanoindentation. The Vickers hardness of SCL LZ was in the range of 3.90–4.22 GPa. On the other hand, nanoindentation showed that the nano-hardness of SCL LZ was in the range of 5.41–6.08 GPa, and the corresponding Young’s modulus was calculated from the nanoindentation data with a value range of 89.0–104.3 GPa. The porosity of the LZ coating has a big influence on the hardness and Young’s modulus. The hardness and Young’s modulus decrease with an increase in the LZ coating porosity. 2. Bond strength test was used to study the adhesion bond strength between the top coat and bond coat. The measured adhesion strength of the as-sprayed SCL LZ coating was about 10.48 MPa, which was lower than the adhesion strength of the SCL 8YSZ coating (13.59 MPa). After thermal spray process, the thermal residual stress was generated in the TBC system. The theoretical thermal residual stress calculation showed that the SCL LZ coating has larger residual stress © Springer Nature Switzerland AG 2021 X. Guo et al., Novel Lanthanum Zirconate-based Thermal Barrier Coatings for Energy Applications, https://doi.org/10.1007/978-3-030-58695-9_6

79

80

6  Concluding Remarks and Future Research Directions

d­ ifference at the interface compared to the 8YSZ coating due to a big CTE difference between LZ coating and the bond coat. In addition, LZ has lower fracture toughness than 8YSZ, indicating that the crack propagation in LZ is much easier than in 8YSZ. The low fracture toughness of LZ also leads to high erosion rate. 3. Thermal conductivity and CTE of porous LZ coating with a porosity of 11.54% were measured using laser flash method and dilatometry method, respectively. The thermal conductivity of porous LZ coating was 0.59–0.68 W/m/oC at a temperature range of 100–900 °C, which was about 25% lower than that of porous 8YSZ coating. CTE values of LZ were approximately 9–10 × 10−6 /K at a temperature range of 400–600 K, which was about 10% lower than that of 8YSZ. The large CTE difference between the LZ and bond coat might lead to large thermal residual stress. 4. The furnace cycling test (FCT), heat treatments, and jet engine thermal shock test (JETS) were performed to investigate the thermal durability performance of LZ-based TBCs. The results showed that all of the layered LZ coatings were delaminated within the first 20 cycles in FCT tests. The large thermal residual stress and low fracture toughens are the primary reasons for LZ coating’s delamination. However, the composite LZ/YSZ TBCs showed a big improvement in the FCT test because the CTE difference between top and bond coats was reduced by the composite coatings and buffer layer. In both heat treatment and the JETS tests, the double-layer coating with porous 8YSZ and LZ layers exhibited a better performance than the other layered LZ-based coating, suggesting that porous 8YSZ can be used as a buffer layer for increasing the lifetime in thermal cycling tests. The double layer composite coating (50% LZ + 50% 8YSZ and 25% LZ + 75% 8YSZ) with 8YSZ buffer layer shows the best thermal durability performance in both FTC and JETS among all the LZ-based TBCs. 5. The thermal gradient mechanical fatigue (TGMF) test was performed to investigate the thermal-mechanical coupled stability. The LZ TBCs samples have fewer cycles than 8YSZ. The large CTE difference between LZ topcoat and bond coat leads to high thermal residual stress, which was accumulated during the TGMF test process. Delamination occurred at the interface of the LZ top coat and bond coat. 6. DFT and MD calculations are effective techniques to investigate the intrinsic physical, mechanical, and thermal properties of various materials. The LZ unit cell was geometrically optimized by minimization of the total energy using DFT calculations. The calculated lattice parameter is 10.875 Å. The nanoscale tensile and shear simulations were conducted based on the optimized LZ unit cell using DFT and MD methods. The elastic modulus, ultimate strength, and toughness were calculated based on the stress-strain curves. The anisotropic Young’s modulus was calculated using two different approaches: (1) the tensile stress-strain relation in large deformation and (2) the elastic constant analysis in elastic deformation. The largest Young modulus is located at [111] direction of LZ unit cell. The anisotropic Young’s modulus was related to the bonds between cation and anion, which were investigated by the electronic charge distribution and charge transfer analysis.

6.2  Outlook of Future Work

81

7. The thermal properties such as specific heat capacity and thermal conductivity were investigated using DFT, MD, and FE methods. The specific heat of LZ unit cell was 0.4–0.46 J/g/K at a temperature range of 673–1673 K, which was computed using DFT calculations. The thermal conductivities of LZ single crystal and porous LZ coating were calculated using multiscale modeling technique, which was based on MD and image-based FE method. The thermal conductivity of single crystal LZ was calculated using MD method, whose value was 1.25–1.39 W/m/K at a temperature range of 473–1273 K. The thermal conductivity of polycrystalline LZ coating was calculated using SEM image-based FE method, whose value was 0.46–0.59  W/m/K at a temperature range of 473–1273 K. The calculated specific heat and thermal conductivity results have a good accuracy, compared with the experimental results. 8. The tensile and shear stress-strain conditions at the interface of top and bond coat have a big influence on the thermal and mechanical durabilities of LZ coating, so the nanoscale tensile and shear simulations are calculated by DFT method. The ZrO2/Ni interface model was chosen as a feasible simple model to mimic the interface between the top and bond coats. Both the Young’s modulus and shear modulus values of ZrO2/Ni interface model were decreased with an increase in the Ni layer thickness, so as the ultimate strength and toughness.

6.2  Outlook of Future Work The physical, mechanical, and thermal properties of lanthanum zirconate-based thermal barrier coatings are systematically discussed in this book. Powder fabrication, deposition technique, microstructure morphology, physical, thermal, and mechanical properties, thermomechanical durability, and failure mechanisms of LZ coating are also discussed. Since the TBC is a quite complicated composite material system, many factors are responsible for the properties and durability of TBCs; for instance, the coating material species, deposition technique, coating porosity, cracks and pore morphology, grain size, dope element choice, coating thickness, heat treatment method, multiple layer coating architecture, and so on. The research on LZ-based TBC is still far from the end. This book discusses several issues of LZ, which need to be solved before LZ can be accepted as a TBC material. LZ possesses a lower fracture toughness. The low fracture toughness leads to the initiation and propagation of micro-cracks even at a very low stress level. The vulnerability of LZ for fracture propagation affects its thermal durability. Therefore, it becomes very important to improve the fracture toughness of LZ-based coatings. Although composite material and graded layered structure help to improve the durability during the thermal cycling test, however in the mean time, additional artificial defects are introduced into the TBC systems, which might reduce the lifetime of the coating. Doping LZ with other elements might be a feasible way to improve its fracture toughness. However, the effects of doping elements and their doping ratio are still unclear, which justify further investigations.

82

6  Concluding Remarks and Future Research Directions

The thermal conductivity can be reduced and the coefficient of thermal expansions can be increased by doping. However, material that has better thermal properties does not guarantee a good thermal cycling performance. The experiments of thermomechanical durability and thermal conductivity measurements need to be conducted. The failure mechanisms of delamination of LZ coatings in thermal cycling, mechanical fatigue cycles, and chemical corrosive environment are not totally clear. A combination of experimental techniques and theoretical modeling tools can be the most effective way to unfold the mystery of failure in the LZ-based TBCs. The modeling studies about TGO initiation and growth, bond coating oxidation, and crack initiation and growth in LZ coating are still challenging work, which will help fully understand those failure mechanisms and improve the design of the TBC coating.

Index

A Adhesion characteristics, 62 Air plasma spray (APS), 5, 18 Archimedes’ principle, 13 Atomistic model, 62 B Bader charge number, 70, 71 Bader method, 62 Bobzin’s thermal cycle experiments, 25 C Charge-charge interaction, 72 Charge density distributions, 70 Charge-optimized many body (COMB3), 72 Coefficient of thermal expansion (CTE), 29 linear, 23 LZ, 23 stress and strain, 24 YSZ TBCs, 24 Constant optimization, 34, 35 Convergent beam electron diffraction (CBED) techniques, 4 Co-precipitation method, 3 D Debye’s theory, 48 Density functional theory (DFT), 29–32, 61 DFT interface modeling adhesion calculations, 62 adhesion energy, 63 atom displacement vectors, 67 deformation strain, 64

interfacial deformation, 69 layer-thickness dependence, 64 shear deformations, 66 shear stress-strain curves, 68 tensile stress-strain curves, 63 Differential scanning calorimetry (DSC), 18 E EB-PVD coating, 18 EB-PVD deposition parameters, 7 Elastic modulus, 65 pulse-echo method, 15 Electron localization function (ELF), 69 metallic bonding, 69 Ni interface models, 69 Erosion, 17 Erosion test performance, 18 F Finite element (FE) method, 53 First principles calculations, 59, 61 Fracture toughness, 16, 17 Furnace cycling test (FCT), 24, 80 G Generalized gradient approximation (GGA), 30, 61 Green-Kubo method, 50 H Hardness test, 16 Heat flux direction, 51

© Springer Nature Switzerland AG 2021 X. Guo et al., Novel Lanthanum Zirconate-based Thermal Barrier Coatings for Energy Applications, https://doi.org/10.1007/978-3-030-58695-9

83

Index

84 I Interface modeling ceramic-metal interface, 59 crystallographic orientation, 60 failure and spallation mechanisms, 59 idealized mechanical properties, 59 toughness and elastic modulus, 60 Interfacial mechanical behaviors, 59 J Jet engine thermal shock (JETS) test, 24, 80 L Lanthanum zirconate (LZ), 2, 79 doping, 81 failure mechanisms, 82 fracture toughness, 81 mechanisms, 81 thermal conductivity, 82 Lanthanum zirconate (LZ) crystal analytical method, 29 anisotropic mechanical properties, 29 bader charge analysis, 40, 41 charge density distribution, 40, 41 constant optimization, 34, 35 DFT, 29–32, 43 elastic constants, 29 elastic moduli, 29 elastic moduli calculation, 37–40 low thermal conductivity, 29 MD model, 33 MD tensile calculations, 41–43 mechanical properties, 29 nanoscale tensile, 29 shear simulations, 41–43 shear stress-strain, 29 stress-strain analysis, 36, 37 thermal cycling performance, 29 Voigt-Reuss-Hill approximation, 44 Lanthanum Zirconate (LZ) powder APS, 6 co-precipitation method, 3 crystal structures, 4 pyrochlore crystal, 4 sol-gel method, 3 X-ray diffraction patterns, 4 XRD pattern, 4 Large-scale Atomic/Molecular Massively Parallel Simulator (LAMMPS), 33 Layered LZ-based TBCs, 79 Lennard-Jones expression, 72 LZ coatings, 9, 24

porosity, 13, 14 properties, 16, 19 sintering, 14 theoretical density, 13 YSZ coating, 16 LZ polycrystalline coating samples, 53 LZ unit cell model, 73 M MD ZrO2(111)/Ni(111) interface model, 74, 75 Melting point, 18 Molecular dynamics (MD), 33, 50, 60 Muller-Plathe algorithm, 50–51 N Non-equilibrium molecular dynamics (NEMD), 50 P Poisson’s ratio, 15 Projector augmented wave (PAW), 30, 61 Pulsed thermal imaging-multilayer analysis (PTI-MLA) method, 47, 55 Pulse-echo method, 15 Q Quantitative imaging analysis method, 15 R Reverse non-equilibrium molecular dynamics (RNEMD) method, 51 S Scanning electron microscope (SEM), 53 Shear deformation model, 62, 65–67 Shear modulus, 60 Shear stress-strain curves, 74 Sintering, 14 behavior, 14 LZ coating, 14 thermomechanical behavior, 14 Sol-gel method, 3 Stress-strain behaviors, 62 atom displacement vectors, 67 tensile models, 63 toughness, 68 Supercell model, 51 Suspension plasma spray (SPS), 8

Index T Temperature distribution, 51 heat flux direction, 52 LZ single crystal, 51 Tensile and shear stress-strain conditions, 81 Tensile deformation, 62, 64 Tensile model, 31 Thermal barrier coatings (TBCs) application, 1 criteria, 2 porosity, 2 structure, 1 surface temperature, 2 Thermal conducitivity, 20–22, 80 MD method, 50 NEMD approach, 50 Thermal conductivity model, 51 Thermal conductivity values, 52 Thermal cycling performance, 25 Thermal cycling tests, 24 Thermal gradient mechanical fatigue (TGMF) test, 80 Thermal properties DSC, 18 experimental methods, 47 melting point, 18 PTI-MLA techniques, 47 SEM microstructures images, 47 TBCs, 47

85 Thermal spray, 3 Thermodynamic energies, 48 curves, 49 entropy, 49 LZ, 49 V van der Waals interactions, 72 Vienna Ab initio Simulation Program (VASP), 30 X X-ray diffraction, 4 Y Young’s modulus, 16, 60, 69, 79 Z ZrO2/Ni ceramic-metal interface, 72 ZrO2 /Ni interface models, 62, 72, 75, 76 Ni atomic layers, 61 plan-wave cutoff energy, 61 symmetric models, 60 tensile stress-strain curves, 74 ZrO2-Ni-ZrO2 sandwich model, 74